Visual Question Answering on 33.3% VQA v2 + 66.7% AdVQA (test)
57.71AccuracyOFA-Large + Selector (Self)
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| OFA-Large + Selector (Self)VQA Model=OFA-Large, VQA Model Train Set=A, Selection Function=Selector, Selection Function Train Set=B, Selection Function Targets=Self2023.06 | 57.71 | 3.03 | 11.2 | 22.04 | 20.44 | 37.45 | -5.27 | -39.01 | |
| OFA-Large + Selector (LYP Variant)VQA Model=OFA-Large, VQA Model Train Set=A, Selection Function=Selector, Selection Function Train Set=A+B, Selection Function Targets=LYP-A Self-B2023.06 | 57.71 | 1.89 | 10.78 | 20.09 | 20.63 | 37.51 | -6.12 | -30.14 | |
| OFA-Large + MaxProbVQA Model=OFA-Large, VQA Model Train Set=A, Selection Function=MaxProb2023.06 | 57.69 | 0.13 | 3.65 | 14.24 | 22.36 | 34.91 | -16.36 | -49.7 | |
| OFA-Large + MaxProbVQA Model=OFA-Large, VQA Model Train Set=A+B, Selection Function=MaxProb2023.06 | 57.52 | 0.08 | 0.54 | 13.41 | 22.87 | 34.7 | -20.37 | -56.21 | |
| OFA-Large + Selector (Self)VQA Model=OFA-Large, VQA Model Train Set=A+B, Selection Function=Selector, Selection Function Train Set=A+B, Selection Function Targets=Self2023.06 | 57.5 | 0.46 | 9.02 | 20.14 | 21.1 | 35.39 | -10.72 | -61.53 | |
| OFA-Large + Selector (LYP)VQA Model=OFA-Large, VQA Model Train Set=A+B, Selection Function=Selector, Selection Function Train Set=A+B, Selection Function Targets=LYP2023.06 | 57.5 | 0.08 | 10.28 | 19.93 | 20.94 | 36.6 | -8.58 | -44.52 | |
| OFA-Base + MaxProbVQA Model=OFA-Base, VQA Model Train Set=A+B, Selection Function=MaxProb2023.06 | 54.28 | 0.03 | 0.53 | 10.16 | 25.72 | 29.96 | -25.56 | -44.75 | |
| OFA-Base + Selector (Self)VQA Model=OFA-Base, VQA Model Train Set=A+B, Selection Function=Selector, Selection Function Train Set=A+B, Selection Function Targets=Self2023.06 | 54.26 | 1.52 | 8.79 | 16.23 | 24.15 | 31.88 | -12.68 | -52.56 | |
| OFA-Base + Selector (LYP)VQA Model=OFA-Base, VQA Model Train Set=A+B, Selection Function=Selector, Selection Function Train Set=A+B, Selection Function Targets=LYP2023.06 | 54.26 | 1.95 | 9.71 | 17.11 | 23.79 | 32.38 | -12.12 | -20.65 | |
| OFA-Base + Selector (Self)VQA Model=OFA-Base, VQA Model Train Set=A, Selection Function=Selector, Selection Function Train Set=B, Selection Function Targets=Self2023.06 | 53.77 | 2 | 8.23 | 15.97 | 24.45 | 30.21 | -10.46 | -16.01 | |
| OFA-Base + Selector (LYP Variant)VQA Model=OFA-Base, VQA Model Train Set=A, Selection Function=Selector, Selection Function Train Set=A+B, Selection Function Targets=LYP-A Self B2023.06 | 53.77 | 1.73 | 7.79 | 15.51 | 24.35 | 30.86 | -7.2 | -28.39 | |
| OFA-Base + MaxProbVQA Model=OFA-Base, VQA Model Train Set=A, Selection Function=MaxProb2023.06 | 53.71 | 0 | 0.45 | 8.44 | 26.47 | 29.64 | -17.6 | -43.15 | |
| CLIP-ViL + MaxProbVQA Model=CLIP-ViL, VQA Model Train Set=A+B, Selection Function=MaxProb2023.06 | 47.94 | 0.67 | 1.28 | 5.59 | 32.08 | 20.87 | -16.85 | -21.99 | |
| CLIP-ViL + Selector (Self)VQA Model=CLIP-ViL, VQA Model Train Set=A+B, Selection Function=Selector, Selection Function Train Set=A+B, Selection Function Targets=Self2023.06 | 47.94 | 0.05 | 1.44 | 5.49 | 31.79 | 22.2 | -11.69 | -15.28 | |
| CLIP-ViL + Selector (LYP)VQA Model=CLIP-ViL, VQA Model Train Set=A+B, Selection Function=Selector, Selection Function Train Set=A+B, Selection Function Targets=LYP2023.06 | 47.94 | 2.13 | 6.6 | 10.44 | 29.77 | 23.6 | -0.77 | -0.89 | |
| CLIP-ViL + MaxProbVQA Model=CLIP-ViL, VQA Model Train Set=A, Selection Function=MaxProb2023.06 | 46.66 | 0 | 0 | 3.04 | 33.67 | 18.32 | -24.68 | -28.56 | |
| CLIP-ViL + Selector (Self)VQA Model=CLIP-ViL, VQA Model Train Set=A, Selection Function=Selector, Selection Function Train Set=B, Selection Function Targets=Self2023.06 | 46.66 | 1.91 | 5.65 | 10.09 | 31.43 | 22 | -11.5 | -12.05 |