ussaema
ussaema
I already looked at the code in models directory, It is CapsuleNet using dynamic routing with margin loss. EM paper has indeed more repeating layers, but the routing (clustering of...
I understand :) you are using the same loss (margin loss) for the generator and the discriminator, right ?
@sukhadj so it worked with keras 2.2.0 for you ?
I am not using this module anymore. I created my own module for GANs for Kreas 2.2.x. check https://github.com/ussaema/Vector_Matrix_CapsuleGAN