Visual Question Answering on RSFG-SC
64.9Scene AccuracyVHM-RL
Evaluation results
| Method | Links | ||
|---|---|---|---|
| VHM-RL2025.11 | 64.9 | 30.89 | |
| RS-EoT-7BBase Model=Qwen2.5-VL-7B2025.11 | 64.05 | 56.52 | |
| WeThinkBase Model=Qwen2.5-VL-7B2025.11 | 60.12 | 38.35 | |
| VL-RBase Model=Qwen2.5-VL-7B2025.11 | 58.82 | 34.84 | |
| Qwen2.5VLBase Model=Qwen2.5-VL-7B2025.11 | 57.42 | 36.78 | |
| MM-EBase Model=Qwen2.5-VL-7B2025.11 | 56.98 | 31.57 | |
| VLAA-TBase Model=Qwen2.5-VL-7B2025.11 | 56.5 | 32.91 | |
| Geo-R12025.11 | 52.46 | 20.82 | |
| R1-OVBase Model=Qwen2.5-VL-7B2025.11 | 47.04 | 24.65 | |
| Vis-R1Base Model=Qwen2.5-VL-7B2025.11 | 21.63 | 6.27 |