Vision-Language Question Answering on VQA v2 (Accuracy)
76.7Accuracy (VQA v2)LoRA
Evaluation results
| Method | Links | |
|---|---|---|
| LoRAModel=Qwen2.5-VL-3B, Optimization=FlatPO2026.06 | 76.7 | |
| LoRAModel=Qwen2.5-VL-3B, Optimization=GAM2026.06 | 74.5 | |
| LoRAModel=Qwen2.5-VL-3B, Optimization=SAM2026.06 | 74.4 | |
| LoRAModel=Qwen2.5-VL-3B, Optimization=Base2026.06 | 73.6 | |
| Prefix TuningModel=Qwen2.5-VL-3B, Optimization=FlatPO2026.06 | 73.6 | |
| Prefix TuningModel=Qwen2.5-VL-3B, Optimization=GAM2026.06 | 72.4 | |
| Prefix TuningModel=Qwen2.5-VL-3B, Optimization=SAM2026.06 | 72.1 | |
| Prefix TuningModel=Qwen2.5-VL-3B, Optimization=Base2026.06 | 70.7 |