Abstract: We propose an alternative generator architecture for generative adversarial networks, borrowing from style transfer literature. The new architecture leads to an automatically learned, ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results