1 paper
Mengyao Cui, Zhe Zhu, Shao-Ping Lu +1
Previous text-to-image synthesis algorithms typically use explicit textual instructions to generate/manipulate images accurately, but they have difficulty adapting to guidance in t…