How to give a custom voice in as speaker?

[fatima-falconit]

the model is working perfectly on predefined speaker ids but I wanna give my own voice for cloning. How can I do that?

[meet1919]

checkout this https://github.com/myshell-ai/OpenVoice

[jidkano]

@meet1919 this seems to be a completely different model that is nowhere near as fast as MeloTTS

[yukiarimo]

You can fine-tune MeloTTS

[rushichavda]

@yukiarimo do you have any resources/code which can help in finetuning? i have 1 hour of english data, and want to add new voice. Thanks

[yukiarimo]

Yes, I'll drop some stuff soon. Prepare you're dataset to be 3-30s chunked audio with transcripts for each.