search

RVC-Boss / GPT-SoVITS

1 min voice data can also be used to train a good TTS model! (few shot voice cloning)
MIT License
35.66k stars 4.07k forks source link

Problem with not being able to generate new sentences or words through voice synthesis #1781

Open kbuwel opened 6 hours ago

kbuwel commented 6 hours ago

This problem occurs in the following situations: The language is Korean. When you run the function to keep the last generated voice, you cannot create new sentences or words. Even if you try to synthesize voice, the previously entered sentences or words are output as voice.

XXXXRT666 commented 4 hours ago

That's why the function means " keep the last generated voice"

kbuwel commented 3 hours ago

I think the translated explanation is insufficient and the functions are insufficient. I wish the voice was output consistently. If the desired tone or intonation is output, I wish it would be maintained. I think I talked about this issue last time, but the answer was to run the function that maintains the output voice. The result of the execution is the problem in the text. If you run this function, you cannot create other sentences or words. If a voice you like is output, you want to maintain this voice and create new sentences or words.