Abstract: Despite extensive research on training generative adversarial networks (GANs) with limited training data, learning to generate images from long-tailed training distributions remains fairly ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results