search

MoonInTheRiver / DiffSinger

DiffSinger: Singing Voice Synthesis via Shallow Diffusion Mechanism (SVS & TTS); AAAI 2022; Official code
MIT License
4.31k stars 713 forks source link

How to train another language? #26

Open Cardroid opened 2 years ago

Cardroid commented 2 years ago

I think this "DiffSinger" model is based on Chinese. Please give me advice on how to train them in another language. Thank you for share!

MoonInTheRiver commented 2 years ago

"DiffSinger" takes in phoneme, pitch, and duration. You need a Grapheme-to-Phoneme tool, like g2p_en for English, pypinyin for Chinese, or something for your language.

Cardroid commented 2 years ago

"DiffSinger" takes in phoneme, pitch, and duration. You need a Grapheme-to-Phoneme tool, like g2p_en for English, pypinyin for Chinese, or something for your language.

Thank you for your answer. I have another question. How do you fine-tune the vocoder for the new dataset?

morganne00 commented 2 years ago

Can you please explain how to train this Diffsinger in a different language? Does the vocoder need to be trained from scratch then, how?

ktroktorin commented 2 years ago

Hi, can you please describe what needs to be done to train Diffsinger in a different language?