Scene Description on WAD (test)
44.5ROUGE-1WalkVLM-LR
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| WalkVLM-LRParameters=2B, Zero-shot=false2025.08 | 44.5 | 31 | 40.6 | 57.6 | 77.6 | |
| Deepseek-VL2-SmallParameters=2.8B, Zero-shot=false2025.08 | 43.1 | 28.5 | 39 | 44.7 | 75.5 | |
| InternVL3-2BParameters=2B, Zero-shot=false2025.08 | 43 | 29 | 39.1 | 45.3 | 64.4 | |
| Ovis2-2BParameters=2B, Zero-shot=false2025.08 | 42.9 | 27.3 | 39.1 | 44.8 | 74.2 | |
| Kimi-VL-A3B-InstructParameters=2.8B, Zero-shot=false2025.08 | 42.8 | 28.9 | 39.3 | 45.9 | 66.2 | |
| Qwen2-VL-2B-InstructParameters=2B, Zero-shot=false2025.08 | 42.2 | 27.9 | 39 | 44.7 | 64.3 | |
| Qwen2.5-VL-3B-InstructParameters=3B, Zero-shot=false2025.08 | 42 | 27.7 | 38.3 | 45.2 | 67.1 | |
| InternVL3-1B-InstructParameters=1B, Zero-shot=false2025.08 | 41.7 | 28 | 38.1 | 43.9 | 63.6 | |
| mPLUG-Owl3-2BParameters=2B, Zero-shot=false2025.08 | 40.1 | 26.4 | 36.6 | 44.5 | 71.2 | |
| GPT-4oZero-shot=true2025.08 | 14.5 | 3.2 | 8.7 | 26.3 | — | |
| Doubao-seed-1-6-flashZero-shot=true2025.08 | 14 | 2.9 | 10.5 | 27 | 53.7 | |
| InternVL3-8B-InstructParameters=8B, Zero-shot=true2025.08 | 11.8 | 3.3 | 9.4 | 29.3 | 41.7 | |
| Gemini-2.5-FlashZero-shot=true2025.08 | 11.7 | 2.4 | 7.9 | 21.6 | 48.9 | |
| Qwen2-VL-72B-InstructParameters=72B, Zero-shot=true2025.08 | 10.7 | 2.8 | 7.9 | 29.3 | 36.3 | |
| Qwen2.5-VL-7B-InstructParameters=7B, Zero-shot=true2025.08 | 10.2 | 2.6 | 7.5 | 24.3 | 37.5 | |
| MiniCPM-V-2-6Parameters=8B, Zero-shot=true2025.08 | 9.1 | 2.6 | 7.2 | 26.1 | 35.1 | |
| MiMo-VL-7B-SFTParameters=7B, Zero-shot=true2025.08 | 8.6 | 2.6 | 6.9 | 25.1 | 43.3 |