1 paper
Huiguo He, Qiuyue Wang, Yuan Zhou +4
Training-free diffusion models have achieved remarkable progress in generating multi-subject consistent images within open-domain scenarios. The key idea of these methods is to inc…