Skip to content

Instantly share code, notes, and snippets.

@myurasov
Last active April 3, 2021 03:12
Show Gist options
  • Save myurasov/6ecf449b32eb263e7d9a7f6e9aed5dc2 to your computer and use it in GitHub Desktop.
Save myurasov/6ecf449b32eb263e7d9a7f6e9aed5dc2 to your computer and use it in GitHub Desktop.
Wasserstein ACGAN in Keras
Display the source blob
Display the rendered blob
Raw
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
@HitLuca
Copy link

HitLuca commented Jun 7, 2018

I assume that your weight initialization is based on the fact that you want your critic to not be almost completely clipped at first iteration. Why then are the weights for the generator also initialized in the same way? Leaving the default initialization should be better for it in my opinion

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment