ICML2025

EasyRef: Omni-Generalized Group Image Reference for Diffusion Models via Multimodal LLM

Zhuofan Zong, Dongzhi Jiang, Bingqi Ma, Guanglu Song, Hao Shao, Dazhong Shen, Yu Liu, Hongsheng Li

摘要

Figure 2. Spatial misalignment issue of the embedding averaging operation. The images with faces are synthetic. methods like LoRA, achieving superior aesthetic quality and robust zero-shot generalization across diverse domains.