Multimodal Evaluation on DID-Bench
41.19CLIP-S ScoreOurs
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| OursModel=GPT-4o2025.04 | 41.19 | 76.98 | 8.42 | 4.3 | |
| SynthesizedModel=GPT-4o2025.04 | 41.13 | 76.87 | 29.71 | 7.09 | |
| w/o. semanticModel=GPT-4o2025.04 | 41.1 | 75.53 | 30.2 | 8.54 | |
| ITModel=GPT-4o2025.04 | 41 | 76.61 | 25.25 | 5.54 | |
| OursModel=LLaVA-1.52025.04 | 40.78 | 72.81 | 49.5 | 15.33 | |
| SynthesizedModel=LLaVA-1.52025.04 | 40.77 | 71.89 | 53.47 | 15.44 | |
| ITModel=LLaVA-1.52025.04 | 40.71 | 71.82 | 52.48 | 15.38 | |
| w/o. semanticModel=LLaVA-1.52025.04 | 40.65 | 71.3 | 81.67 | 23.89 | |
| GlobalModel=GPT-4o2025.04 | 40.59 | 76.35 | 11.18 | 5.59 | |
| GlobalModel=LLaVA-1.52025.04 | 40.59 | 71.42 | 52.48 | 17.41 | |
| PoCaModel=GPT-4o2025.04 | 38.36 | 72.47 | 18.32 | 7.55 | |
| PoCaModel=LLaVA-1.52025.04 | 37.26 | 66.64 | 55.45 | 27.5 |