Can I use my own audio data to train the model?

[Sartner]

Can I use my own audio data to train the model? I hope to use my own voice as output.

[notune]

Here is an attempt to implement that, but it's not working really well yet https://github.com/serp-ai/bark-with-voice-clone

[theguaz]

@Noahs-Git do you have an idea of what's wrong, I'm looking for a way to do voice cloning and I hav only found Tacotron as an alternative so far

Cheerios!

[gkucsko]

thanks yeah, also see here: https://github.com/suno-ai/bark/issues/37

[mrbb]

Thank you, I have tried it and the result is not very satisfactory. Currently, it is not suitable for practical use.