Open szy96 opened 1 year ago
Hi, thanks for your kind words.
I would add it the same way rosinality does for the style in that ModulatedConv2d
module. You could experiment combining the class and the activation with element-wise multiplication or other operators.
Hi thanks for your great job! This is such the work I am always looking for. But I am wondering how I can upgrade the styleGAN in your architecture to styleGAN2, based on https://github.com/rosinality/stylegan2-pytorch.
Thanks a lot!