#wgans — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #wgans, aggregated by home.social.
-
Following up on this, I also explored a more direct use of #WassersteinDistance in #WGANs: Instead of training a discriminator, the generator is optimized by explicitly computing the #OptimalTransport distance between real and generated samples. This turns the loss into the actual metric of interest and removes the adversarial setup, leading to a more direct and stable training signal. And we can generate cool animations, too ^_^
🌍 https://www.fabriziomusacchio.com/blog/2023-07-30-wgan_with_direct_wasserstein_distance/
-
Eliminating the middleman: You can apply the computation of the #Wasserstein distance even more directly in #WassersteinGANs (#WGANs), eliminating the need for a discriminator.
🌎 https://www.fabriziomusacchio.com/blog/2023-07-30-wgan_with_direct_wasserstein_distance/