Object Hallucination Mitigation on MSCOCO 2014 (val)
32.4CHAIR Specificity ScoreGreedy
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| GreedyBackbone=MiniGPT-42024.12 | 32.4 | 12.2 | 14.57 | |
| DoLaBackbone=MiniGPT-42024.12 | 31.9 | 12.15 | 14.54 | |
| Beam SearchBackbone=MiniGPT-42024.12 | 30.1 | 11.87 | 15.35 | |
| OPERABackbone=MiniGPT-42024.12 | 29.7 | 11.96 | 14.82 | |
| VCDBackbone=MiniGPT-42024.12 | 29 | 12.64 | 14.42 | |
| WoodpeckerBackbone=MiniGPT-42024.12 | 28.87 | 10.2 | 15.3 | |
| LUREBackbone=MiniGPT-42024.12 | 27.88 | 10.2 | 15.03 | |
| WoodpeckerBackbone=mPLUG-Owl22024.12 | 26.33 | 8.43 | 16.43 | |
| HALCBackbone=MiniGPT-42024.12 | 25.2 | 9.42 | 14.91 | |
| WoodpeckerBackbone=LLaVA-1.52024.12 | 23.85 | 7.5 | 17.05 | |
| GreedyBackbone=mPLUG-Owl22024.12 | 22.9 | 8.62 | 15.01 | |
| VCDBackbone=mPLUG-Owl22024.12 | 22.8 | 8.68 | 15.14 | |
| DoLaBackbone=mPLUG-Owl22024.12 | 22.4 | 8.36 | 15.13 | |
| NulluBackbone=MiniGPT-42024.12 | 21.4 | 8.99 | 14.81 | |
| LUREBackbone=mPLUG-Owl22024.12 | 21.27 | 7.67 | 15.65 | |
| GreedyBackbone=LLaVA-1.52024.12 | 20.4 | 7.08 | 15.72 | |
| Beam SearchBackbone=mPLUG-Owl22024.12 | 20.3 | 7.62 | 15.43 | |
| VCDBackbone=LLaVA-1.52024.12 | 20.3 | 7.28 | 14.53 | |
| DoLaBackbone=LLaVA-1.52024.12 | 20.2 | 6.75 | 15.68 | |
| OPERABackbone=mPLUG-Owl22024.12 | 20.07 | 7.18 | 15.41 | |
| Beam SearchBackbone=LLaVA-1.52024.12 | 19.5 | 6.84 | 15.99 | |
| LUREBackbone=LLaVA-1.52024.12 | 19.48 | 6.5 | 15.97 | |
| HALCBackbone=mPLUG-Owl22024.12 | 18.8 | 7 | 15.33 | |
| OPERABackbone=LLaVA-1.52024.12 | 17.5 | 6.07 | 16.02 | |
| HALCBackbone=LLaVA-1.52024.12 | 16.9 | 5.72 | 16.02 | |
| NulluBackbone=mPLUG-Owl22024.12 | 15.6 | 5.77 | 15.45 | |
| NulluBackbone=LLaVA-1.52024.12 | 15.2 | 5.3 | 15.69 |