🦖 TEMPFLOW-GRPO: WHEN TIMING MATTERS FOR GRPO IN FLOW MODELS

TempFlow-GRPO (Temporal Flow GRPO), a principled GRPO framework that captures and exploits the temporal structure inherent in flow-based generation.

Paper Code
TempFlow-GRPO Structure TempFlow-GRPO Teaser

🗺️ Roadmap for TempFlow-GRPO

TempFlow-GRPO (Temporal Flow GRPO), a principled GRPO framework that captures andexploits the temporal structure inherent in flow-based generation. TempFlow-GRPO introduces two key innovations: (i) a trajectory branching mechanism that provides process rewards by concentrating stochasticity at designated branching points, enabling precise credit assignment without requiring specialized intermediate reward models; and (ii) a noise-aware weighting scheme that modulates policy optimization according to the intrinsic exploration potential of each timestep, prioritizing learning during high-impact early stages while ensuring stable refinement in later phases. These innovations endow the model with temporally-aware optimization that respects the underlying generative dynamics, leading to state-of-the-art performance in human preference alignment and standard text-to-image benchmark.

Welcome Ideas and Contribution. Stay tuned!

🆕 News

We have presented an improved Flow-GRPO method, TempFlow-GRPO. We will release our code soon!🔥🔥🔥

  • [2025-08-06] We have released the first version of our paper. 🔥🔥🔥
  • [2025-08-11] Thanks Jie Liu's comments for our paper. We will release the 1024 Flux RL model in the month. 🔥🔥🔥
  • [2025-08-14] Our method also achieves better performance in FLUX 1024px with HPSv3 (based on Qwen2-VL) as reward, blue is TempFlow-GRPO and Purple is Flow-GRPO Fixed. 🔥🔥🔥
  • [2025-08-20] We have released the first version of our paper on Hugging Face. 🔥🔥🔥
  • [2025-09-12] We will release the second version of our paper in next week. 🔥🔥🔥
  • [2025-09-17] We will release the code of our paper. 🔥🔥🔥

🚀 Updates

To support research and the open-source community, we will release the entire project—including datasets, training pipelines, and model weights. Thank you for your patience and continued support! ☀️

  • Release arXiv paper
  • Release GitHub repo
  • Release training code
  • Release neat training code
  • Release model checkpoints

📊 Experimental Performance

PickScore

📺 Visualization

Appendix Visualization
Appendix Visualization
Appendix Visualization 2
Appendix Visualization

For more details please read our paper.
Download

BibTeX

@misc{2508.04324,
    Author = {Xiaoxuan He and Siming Fu and Yuke Zhao and Wanli Li and Jian Yang and Dacheng Yin and Fengyun Rao and Bo Zhang},
    Title = {TempFlow-GRPO: When Timing Matters for GRPO in Flow Models},
    Year = {2025},
    Eprint = {arXiv:2508.04324},
}