Open-ended Generation on COCO 2014 (val)
8.58AccuracyLogicCheckGPT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| LogicCheckGPTBase Model=QWEN-VL-Chat2024.02 | 8.58 | 9.96 | |
| vanillaBase Model=QWEN-VL-Chat2024.02 | 8.36 | 9.96 | |
| LogicCheckGPTBase Model=LLaVA-1.52024.02 | 6.5 | 7.64 | |
| LogicCheckGPTBase Model=MiniGPT-42024.02 | 6.02 | 8.38 | |
| vanillaBase Model=LLaVA-1.52024.02 | 5.22 | 7.24 | |
| vanillaBase Model=MiniGPT-42024.02 | 5 | 7.96 | |
| LogicCheckGPTBase Model=mPLUG-Owl2024.02 | 4.32 | 8.74 | |
| vanillaBase Model=mPLUG-Owl2024.02 | 3.44 | 8.78 |