DCPO-T2I

Welcome to the DCPO (Dual Caption Preference Optimization) organization! We're the authors of the Pick-Double Caption dataset. We have hosted the dataset, models trained on the dataset, and other artifacts under this organization. You will be able to find them below. Additionally, here are the links to the paper and the project website:

• arXiv

• Paper

• Codebase

  Update October 23, 2025: We got accepted to TMLR journal 🥳