Debiasing Pretrained Generative Models by Uniformly Sampling Semantic Attributes

Open in new window