Medical Visual Question Answering on OBScan Overall
92.13AccuracyPhi3.5V-Med
Evaluation results
| Method | Links | |
|---|---|---|
| Phi3.5V-MedEvaluation Protocol=Fine-Tuning, Decoding Strategy=ARCD2025.12 | 92.13 | |
| Phi3.5V-MedEvaluation Protocol=Fine-Tuning, Decoding Strategy=Greedy2025.12 | 89.76 | |
| Phi3.5V-MedEvaluation Protocol=Fine-Tuning, Decoding Strategy=VCD2025.12 | 89.76 | |
| Phi3.5V-MedEvaluation Protocol=Fine-Tuning, Decoding Strategy=DoLA2025.12 | 89.37 | |
| Phi3.5V-MedEvaluation Protocol=Fine-Tuning, Decoding Strategy=OPERA2025.12 | 87.4 | |
| GPT-4oEvaluation Protocol=General Domain, Decoding Strategy=Greedy2025.12 | 71.26 | |
| Phi3.5V-MedEvaluation Protocol=Zero-Shot, Decoding Strategy=ARCD2025.12 | 65.75 | |
| HuatuoV-34BEvaluation Protocol=Zero-Shot, Decoding Strategy=Greedy2025.12 | 64.96 | |
| Qwen-VL-7BEvaluation Protocol=General Domain, Decoding Strategy=Greedy2025.12 | 63.78 | |
| Phi3.5V-MedEvaluation Protocol=Zero-Shot, Decoding Strategy=VCD2025.12 | 62.99 | |
| Phi3.5V-MedEvaluation Protocol=Zero-Shot, Decoding Strategy=OPERA2025.12 | 62.06 | |
| Phi3.5V-MedEvaluation Protocol=Zero-Shot, Decoding Strategy=Greedy2025.12 | 61.81 | |
| Phi3.5V-MedEvaluation Protocol=Zero-Shot, Decoding Strategy=DoLA2025.12 | 61.02 | |
| LLaVA-1.6-7BEvaluation Protocol=General Domain, Decoding Strategy=Greedy2025.12 | 57.48 | |
| LLaVA-Med-7BEvaluation Protocol=Zero-Shot, Decoding Strategy=Greedy2025.12 | 57.48 | |
| Phi3.5V-4.2BEvaluation Protocol=General Domain, Decoding Strategy=Greedy2025.12 | 50.39 |