UDM-GRPO: Stable and Efficient Group Relative Policy Optimization for Uniform Discrete Diffusion Models 文章

ArXiv CS.CV2026-05-28NEWSen作者: Jiaqi Wang, Haoge Deng, Ting Pan, Yang Liu, Chengyuan Wang, Fan Zhang, Yonggang Qi, Xinlong Wang

查看原文 →

UDM-GRPO: Stable and Efficient Group Relative Policy Optimization for Uniform Discrete Diffusion Models · 相关技术

相关技术

RL CFG-Free Reduced-Step reinforcement learning T2I UDM-GRPO GRPO Uniform Discrete Diffusion Model