1 paper
Zhihan Chen, Yuhuan Zhao, Yijie Zhu +1
Subject-driven text-to-image diffusion models have achieved remarkable success in preserving single identities, yet their ability to compose multiple interacting subjects remains l…