Diffusion CoT

This organization holds the artifacts for our research conducted on enabling reasoning in diffusion-based image synthesis models. Our first effort in this line of research is ReflectionFlow, where we introduce the first ever large-scale dataset, GenRef, suitable for reflection-tuning.

Below, we provide the links related to ReflectionFlow:

• ReflectionFlow paper
• Projection website
• Models and datasets
• Code

Citation

misc{zhuo2025reflectionperfectionscalinginferencetime,
      title={From Reflection to Perfection: Scaling Inference-Time Optimization for Text-to-Image Diffusion Models via Reflection Tuning}, 
      author={Le Zhuo and Liangbing Zhao and Sayak Paul and Yue Liao and Renrui Zhang and Yi Xin and Peng Gao and Mohamed Elhoseiny and Hongsheng Li},
      year={2025},
      eprint={2504.16080},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2504.16080}, 
}

Enjoy 🤗