Visual Question Answering on GQA (Orig.)
58.35Overall AccuracyMCAN-Small
Evaluation Results
| Method | Links | |
|---|---|---|
| MCAN-SmallFaster R-CNN bbox features=true, Backbone=ResNet-1012022.04 | 58.35 | |
| RelViTFaster R-CNN bbox features=false, Backbone=PVTv2-b22022.04 | 57.87 | |
| EsViTFaster R-CNN bbox features=false, Backbone=PVTv2-b22022.04 | 56.95 | |
| ViT-onlyFaster R-CNN bbox features=false, Backbone=PVTv2-b22022.04 | 56.62 | |
| MACFaster R-CNN bbox features=true, Backbone=ResNet-1012022.04 | 54.06 | |
| BottomUpFaster R-CNN bbox features=true, Backbone=ResNet-1012022.04 | 53.21 | |
| MCAN-SmallFaster R-CNN bbox features=false, Backbone=ResNet-1012022.04 | 51.1 |