RubricRL: Simple Generalizable Rewards for Text-to-Image Generation
Open paper
Benchmark
Model
Rank
Results
text-to-image-generation-on-dpg-bench
RubricRL (Phi3-3.8B) (extra data)
–
DPG-Bench Score: 86.07
text-to-image-generation-on-geneval
RubricRL (Phi3-3.8B) (extra data)
–
Overall: 0.8468
Rank counts only results with a code link.