Object Presence Detection on POPE average (MS-COCO, A-OKVQA, GQA)
89.21AccuracyReVisiT
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ReVisiTBackbone=InternVL3-8B2025.06 | 89.21 | 88.63 | 89.44 | |
| GreedyBackbone=InternVL3-8B2025.06 | 89.17 | 88.87 | 89.36 | |
| M3IDBackbone=InternVL3-8B2025.06 | 89.05 | 88.87 | 89.22 | |
| CODEBackbone=InternVL3-8B2025.06 | 86.7 | 85.4 | 87.22 | |
| DoLaBackbone=InternVL3-8B2025.06 | 86.63 | 92.08 | 85.64 | |
| SIDBackbone=Qwen2.5-VL-7B2025.06 | 85.78 | 92.27 | 84.72 | |
| VCDBackbone=InternVL3-8B2025.06 | 85.39 | 86.33 | 85.41 | |
| SIDBackbone=InternVL3-8B2025.06 | 85.38 | 86.46 | 85.44 | |
| ReVisiTBackbone=Qwen2.5-VL-7B2025.06 | 84.53 | 93 | 82.93 | |
| VCDBackbone=Qwen2.5-VL-7B2025.06 | 84.43 | 92.48 | 82.91 | |
| GreedyBackbone=Qwen2.5-VL-7B2025.06 | 84.23 | 93.01 | 82.53 | |
| M3IDBackbone=Qwen2.5-VL-7B2025.06 | 83.56 | 91.52 | 82 | |
| CODEBackbone=Qwen2.5-VL-7B2025.06 | 83.55 | 90.2 | 82.26 | |
| ReVisiTBackbone=LLaVA-1.5-7B2025.06 | 81.8 | 78.03 | 83.45 | |
| DoLaBackbone=Qwen2.5-VL-7B2025.06 | 81.36 | 92.36 | 78.71 | |
| SIDBackbone=LLaVA-1.5-7B2025.06 | 80.4 | 75.94 | 82.91 | |
| DoLaBackbone=LLaVA-1.5-7B2025.06 | 79.58 | 74.67 | 82.41 | |
| M3IDBackbone=LLaVA-1.5-7B2025.06 | 79.48 | 74.48 | 82.37 | |
| GreedyBackbone=LLaVA-1.5-7B2025.06 | 79.47 | 74.48 | 82.36 | |
| CODEBackbone=LLaVA-1.5-7B2025.06 | 78.58 | 73.98 | 81.35 | |
| VCDBackbone=LLaVA-1.5-7B2025.06 | 77.74 | 72.51 | 80.97 |