Aligning Few-Step Diffusion Models with Dense Reward Difference Learning

Re-assign community