shentao-yang/dense_reward_t2i

Source code for "A Dense Reward View on Aligning Text-to-Image Diffusion with Preference" (ICML'24).

dense_reward_t2i's Issues

Human evaluation data

Hey, thanks for your great work and releasing it publicly.
I am quite interested in the human evaluation data that you collected (Table 3, Appendix D.3). Are you planning to release the exact prompts used, the images used in each binary comparison, and the human preferences for each sample pair? This would be very helpful!