1 paper
Zilai Zeng, Mingdeng Cao, Zijie Li +5
Recent advances in visual generative models have enabled high-fidelity image editing guided by human instructions. However, these models often struggle with complex instructions in…