probability-recalibration

Tag

Cards List
#probability-recalibration

ReNFT: Repairing Mode Collapse in Reward Post-Training via Internal Probability-Mass Recalibration

arXiv cs.LG · 3d ago Cached

ReNFT is a method to repair mode collapse in reward post-training of diffusion generators by internally recalibrating probability mass, improving diversity while retaining high reward scores.

0 favorites 0 likes
← Back to home

Submit Feedback