autoencoding_beyond_pixels
autoencoding_beyond_pixels copied to clipboard
About trend of reconstruction loss?
Thank you very much for your code! I have a question about the reconstruction loss. I'm a novice and only use pytorch. I've done experiments with other pytorch codes and found that the reconstruction loss of my code is basically reduced. I think my generation result is average. I want to ask if the reduction of reconstruction loss is the right phenomenon for vaegan?