1 paper
Liyu Jia, Fengda Zhang, Jiachun Pan +7
Incorporating visual semantic representations as an intermediate step before image generation can reduce the modeling difficulty between text and images, thereby improving generati…