1 paper
Siwei Xia, Xueqi Hu, Li Sun +1
In StyleGAN, convolution kernels are shaped by both static parameters shared across images and dynamic modulation factors w+∈W+ specific to each image. Therefore, $…