Zhang Zhang

E-GRPO: High Entropy Steps Drive Effective Reinforcement Learning for Flow Models

By Shengjun Zhang