Image Captioning on Flickr-30K (AURC Metrics)
0.253AURC (CIDEr-N)ReCoVERR
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ReCoVERR2026.01 | 0.253 | 0.327 | |
| SeeTRUE2026.01 | 0.251 | 0.411 | |
| VQAScore2026.01 | 0.241 | 0.312 | |
| PAPSP-SModel Scale=Small, Memory Augmentation=false2026.01 | 0.237 | 0.311 | |
| PAPSPBackbone=SigLIP B/162026.01 | 0.237 | 0.311 | |
| LYP2026.01 | 0.236 | 0.323 | |
| MA-PAPSP-SModel Scale=Small, Memory Augmentation=true2026.01 | 0.235 | 0.31 | |
| MA-PAPSPBackbone=SigLIP B/162026.01 | 0.235 | 0.31 | |
| Qwen-VL-7BMethod Setup=LVLM-PAPSP, Backbone Model=Qwen-VL-7B2026.01 | 0.233 | 0.312 | |
| PAPSP-LModel Scale=Large, Memory Augmentation=false2026.01 | 0.229 | 0.307 | |
| PAPSPBackbone=SigLIP SO-400M2026.01 | 0.229 | 0.307 | |
| Qwen-VL-72BMethod Setup=LVLM-PAPSP, Backbone Model=Qwen-VL-72B2026.01 | 0.227 | 0.305 | |
| MA-PAPSP-LModel Scale=Large, Memory Augmentation=true2026.01 | 0.219 | 0.297 | |
| MA-PAPSPMethod Setup=MA-PAPSP, Backbone Model=SigLIP SO-400M2026.01 | 0.219 | 0.297 | |
| MA-PAPSPBackbone=SigLIP SO-400M2026.01 | 0.219 | 0.297 |