search

auspicious3000 / autovc

AutoVC: Zero-Shot Voice Style Transfer with Only Autoencoder Loss
https://arxiv.org/abs/1905.05879
MIT License
976 stars 207 forks source link

Number of speakers used to train downloadable "pre-trained model" #117

Closed MelissaChen15 closed 1 year ago

MelissaChen15 commented 1 year ago

Hi,

Could you please tell us what is the dataset and the number of speakers used in training the pre-trained model available on the readme page?

Thank you!

auspicious3000 commented 1 year ago

20 speakers from the VCTK dataset.

MelissaChen15 commented 1 year ago

Thank you!