Visual Question Answering on VQA-CE Easy
80.5AccuracyVilBERT
Evaluation Results
| Method | Links | |
|---|---|---|
| VilBERTpretrained=Conceptual Caption, fine-tuned=VQA v22021.04 | 80.5 | |
| BLOCK2021.04 | 77.65 | |
| UpDown2021.04 | 76.69 | |
| LfFBase architecture=UpDown [3]2021.04 | 76.6 | |
| RandImgBase architecture=UpDown [3]2021.04 | 76.21 | |
| ESRBase architecture=UpDown [3]2021.04 | 76.18 | |
| RUBiBase architecture=UpDown [3]2021.04 | 75.03 | |
| LMH + RMFEBase architecture=UpDown [3]2021.04 | 73.32 | |
| LMHBase architecture=UpDown [3]2021.04 | 73.12 | |
| SANfeatures=grid2021.04 | 68.45 | |
| LMH+CSSBase architecture=UpDown [3]2021.04 | 62.08 | |
| Shortcuts2021.04 | 61.13 | |
| Question-Only2021.04 | 58.61 | |
| Image-Only2021.04 | 33.58 |