1 paper · 1 filter
Xiangwei Shen, Zhimin Li, Zhantao Yang +6
Recent studies have demonstrated the effectiveness of directly aligning diffusion models with human preferences using differentiable reward. However, they exhibit two primary chall…