Expert Reasoning on SimpleVQA
39.78AccuracyMaLoRA
Evaluation results
| Method | Links | |
|---|---|---|
| MaLoRAModel=Qwen3-VL-8B, Training examples=1.8k2025.10 | 39.78 | |
| MaLoRAModel=Qwen2.5-VL-7B, Training examples=1.8k2025.10 | 39.33 | |
| LoRAModel=Qwen3-VL-8B, Training examples=1.8k2025.10 | 38.65 | |
| LoRAModel=Qwen2.5-VL-7B, Training examples=1.8k2025.10 | 37.98 | |
| BaseModel=Qwen3-VL-8B, Training examples=1.8k2025.10 | 37.98 | |
| QLoRAModel=Qwen3-VL-8B, Training examples=1.8k2025.10 | 36.63 | |
| BaseModel=Qwen2.5-VL-7B, Training examples=1.8k2025.10 | 35.51 | |
| QLoRAModel=Qwen2.5-VL-7B, Training examples=1.8k2025.10 | 35.51 | |
| MaLoRAModel=LLaVA-1.5-7B, Training examples=1.8k2025.10 | 19.78 | |
| QLoRAModel=LLaVA-1.5-7B, Training examples=1.8k2025.10 | 19.32 | |
| LoRAModel=LLaVA-1.5-7B, Training examples=1.8k2025.10 | 18.2 | |
| BaseModel=LLaVA-1.5-7B, Training examples=1.8k2025.10 | 14.16 |