Multimodal Understanding on POPE (test)
87.3AccuracySpatial Relation
Evaluation Results
| Method | Links | |
|---|---|---|
| Spatial RelationTrain (und)=Spatial Relation2026.07 | 87.3 | |
| BaselineTrain (und)=Baseline2026.07 | 86.7 | |
| Text recognitionTrain (und)=Text recognition2026.07 | 86.7 | |
| CountingTrain (und)=Counting2026.07 | 86.4 |