Object Hallucination Evaluation on COCO (val)
88.57Random AccuracyInstructBLIP
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| InstructBLIPLLM Size=13B, Zero-shot=true2023.11 | 88.57 | 89.27 | 56.57 | 82.77 | 84.66 | 62.37 | 72.1 | 77.32 | 73.03 | |
| Chat-UniViLLM Size=7B, Zero-shot=true, multi-scale=true2023.11 | 85.19 | 86.05 | 54.67 | 69.5 | 74.39 | 69.1 | 64.97 | 71.54 | 73.1 | |
| MiniGPT-4LLM Size=13B, Zero-shot=true2023.11 | 79.67 | 80.17 | 52.53 | 69.73 | 73.02 | 62.2 | 65.17 | 70.42 | 67.77 | |
| Chat-UniViLLM Size=7B, Zero-shot=true, multi-scale=false2023.11 | 73.88 | 79.3 | 74.63 | 56.36 | 69.01 | 90.83 | 55.63 | 68.67 | 91.63 | |
| LLaVALLM Size=7B, Zero-shot=true, implementation=re-implementation by current paper authors2023.11 | 72.16 | 78.22 | 76.29 | 61.37 | 71.52 | 85.63 | 58.67 | 70.12 | 88.33 | |
| LLaVALLM Size=13B, Zero-shot=true2023.11 | 64.12 | 73.38 | 83.26 | 63.9 | 72.63 | 81.93 | 58.91 | 69.95 | 86.76 | |
| mPLUG-OwlLLM Size=7B, Zero-shot=true2023.11 | 53.97 | 68.39 | 95.63 | 50.9 | 66.94 | 98.57 | 50.67 | 66.82 | 98.67 | |
| MultiModal-GPTLLM Size=7B, Zero-shot=true2023.11 | 50.1 | 66.71 | 99.9 | 50 | 66.67 | 100 | 50 | 66.67 | 100 |