Categorical Visual Question Answering on A-OKVQA
83AccuracySN
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SNBase Model=Qwen-VL-4b-Instruct, Probing Set Size=3,000, Configuration=Best probed2026.03 | 83 | 82.6 | |
| Qwen-VL-4b-InstructProbing Set Size=3,0002026.03 | 82.7 | 82.3 | |
| SNBase Model=LLaVA-v1.5-7b, Probing Set Size=3,000, Configuration=Best probed2026.03 | 68 | 73.8 | |
| LLaVA-v1.5-7bProbing Set Size=3,0002026.03 | 67.8 | 74.5 |