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

downsample factor #98

Closed inconnu11 closed 2 years ago

inconnu11 commented 2 years ago

Hi, in you code , you downsample the latent representation with self.freq set as 16. I don't understand why the down/up sample factor is 32 as described in your paper.

auspicious3000 commented 2 years ago

16 is a good starting point