Making zero-shot model - Githubissues

auspicious3000 / autovc

AutoVC: Zero-Shot Voice Style Transfer with Only Autoencoder Loss

https://arxiv.org/abs/1905.05879

MIT License

983 stars 207 forks source link

Making zero-shot model #53

Open sbkim052 opened 3 years ago

sbkim052 commented 3 years ago

Thank you for sharing your work.

When i am to make zero shot model, should i train speaker embedder as well as the conversion model with large dataset (VCTK)? Or is it ok to only train the conversion model with VCTK?

ruclion commented 3 years ago

Thank you for sharing your work.

When i am to make zero shot model, should i train speaker embedder as well as the conversion model with large dataset (VCTK)? Or is it ok to only train the conversion model with VCTK?

so how do you do finally? I think if we have large dataset, we have lots of choose

train speaker encoder, then train autovc
only train autovc
train nothing, just conversion it

But I don't know which one is the best....