Question Answering on HiTab (held-in)
58.2AccuracyQwen2-VL-7B-Table-R1
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2-VL-7B-Table-R1Resolution=Dyn.2025.09 | 58.2 | |
| Qwen2-VL-7B-GRPOResolution=Dyn.2025.09 | 54.7 | |
| Qwen2-VL-7B-SFTResolution=Dyn.2025.09 | 48.5 | |
| Gemini-2.5-ProResolution=UNK2025.09 | 46.44 | |
| QVQ-72B-PreviewResolution=Dyn.2025.09 | 45.7 | |
| OpenAI-o4-miniResolution=UNK2025.09 | 44.8 | |
| Qwen2-VL-72BResolution=Dyn.2025.09 | 44.04 | |
| Claude-3.5-SonnetResolution=UNK2025.09 | 41.31 | |
| GPT-4oResolution=UNK2025.09 | 39.32 | |
| Qwen2-VL-2B-Table-R1Resolution=Dyn.2025.09 | 37.3 | |
| Qwen2-VL-2B-GRPOResolution=Dyn.2025.09 | 35.2 | |
| Qwen2-VL-7BResolution=Dyn.2025.09 | 33.1 | |
| Qwen2-VL-2B-SFTResolution=Dyn.2025.09 | 31.9 | |
| Qwen2-VL-2BResolution=Dyn.2025.09 | 22.9 | |
| DeepSeek-VL2 4.5BResolution=Dyn.2025.09 | 18.89 | |
| Table-LLaVA 13BResolution=3362025.09 | 10.85 | |
| Table-LLaVA 7BResolution=3362025.09 | 10.09 | |
| Monkey 7BResolution=8962025.09 | 6.41 | |
| Qwen2-VL-7B-InsResolution=Dyn2025.09 | 5.33 | |
| LLaVA-v1.5 7BResolution=3362025.09 | 2.03 | |
| mPLUG-Owl2 7BResolution=4482025.09 | 0.13 |