1 paper
Kyungmin Lee, Xiaohang Li, Qifei Wang +7
Aligning text-to-image (T2I) diffusion models with preference optimization is valuable for human-annotated datasets, but the heavy cost of manual data collection limits scalability…