Text-to-Image Generation on Affordance-aware Generative Design Evaluation Set
4.08FaithfulnessDALL-E
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| DALL-EEvaluator=GPT-4o-mini2025.02 | 4.08 | 4.32 | 3.32 | 4.56 | |
| SYNTHIAEvaluator=GPT-4o-mini2025.02 | 4.01 | 4.57 | 3.41 | 4.78 | |
| SYNTHIA2025.02 | 3.81 | 3.89 | 3.38 | 4.06 | |
| Stable DiffusionEvaluator=GPT-4o-mini2025.02 | 3.71 | 3.78 | 3.37 | 3.37 | |
| ConceptLabEvaluator=GPT-4o-mini2025.02 | 3.33 | 4.57 | 3.03 | 4.04 | |
| Kandinsky3Evaluator=GPT-4o-mini2025.02 | 3.28 | 4.07 | 3.08 | 4.08 | |
| Stable Diffusion2025.02 | 2.96 | 2.44 | 3.02 | 2.75 | |
| Kandinsky32025.02 | 2.95 | 2.98 | 3.01 | 3.41 | |
| ConceptLab2025.02 | 2.73 | 3.11 | 2.68 | 3.54 |