You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Apr 30, 2020. It is now read-only.
The implementation of VAE with GAN learned similarity metric is sub-optimal. I conducted experiments recently and obtained a stable solution that performs much better. Please see the same model on https://github.com/HitLuca/GAN-collection to check how the VAE loss is improved.
The text was updated successfully, but these errors were encountered:
HitLuca
changed the title
VAE-GAN Generator loss implementation error
VAE-GAN Generator loss suboptimal
Jul 26, 2019
Sign up for freeto subscribe to this conversation on GitHub.
Already have an account?
Sign in.
The implementation of VAE with GAN learned similarity metric is sub-optimal. I conducted experiments recently and obtained a stable solution that performs much better. Please see the same model on https://github.com/HitLuca/GAN-collection to check how the VAE loss is improved.
The text was updated successfully, but these errors were encountered: