VisionDirector: Vision-Language Guided Closed-Loop Refinement for Generative Image Synthesis
Open paper
Benchmark
Model
Rank
Results
image-editing-on-imgedit-data
VisionDirector
–
Overall: 4.35
Add: 4.55
Adjust: 4.32
Extract: 3.72
Replace: 4.83
…
text-to-image-generation-on-geneval
VisionDirector
–
Overall: 0.94
Rank counts only results with a code link.