Image Captioning on MS-COCO (AURC Metrics)
0.158AURC (CIDEr-N)SeeTRUE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SeeTRUE2026.01 | 0.158 | 0.316 | |
| LYP2026.01 | 0.152 | 0.342 | |
| VQAScore2026.01 | 0.146 | 0.321 | |
| ReCoVERR2026.01 | 0.142 | 0.336 | |
| PAPSP-SModel Scale=Small, Memory Augmentation=false2026.01 | 0.142 | 0.317 | |
| PAPSPBackbone=SigLIP B/162026.01 | 0.142 | 0.317 | |
| PAPSP-LModel Scale=Large, Memory Augmentation=false2026.01 | 0.136 | 0.311 | |
| PAPSPBackbone=SigLIP SO-400M2026.01 | 0.136 | 0.311 | |
| MA-PAPSP-SModel Scale=Small, Memory Augmentation=true2026.01 | 0.121 | 0.312 | |
| MA-PAPSPBackbone=SigLIP B/162026.01 | 0.121 | 0.312 | |
| Qwen-VL-7BMethod Setup=LVLM-PAPSP, Backbone Model=Qwen-VL-7B2026.01 | 0.12 | 0.301 | |
| Qwen-VL-72BMethod Setup=LVLM-PAPSP, Backbone Model=Qwen-VL-72B2026.01 | 0.115 | 0.294 | |
| MA-PAPSP-LModel Scale=Large, Memory Augmentation=true2026.01 | 0.109 | 0.286 | |
| MA-PAPSPMethod Setup=MA-PAPSP, Backbone Model=SigLIP SO-400M2026.01 | 0.109 | 0.286 | |
| MA-PAPSPBackbone=SigLIP SO-400M2026.01 | 0.109 | 0.286 |