Object Detection on V3Det (val)
35.4APbAPA*+AGLU-CascadeRCNN-ResNet50
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| APA*+AGLU-CascadeRCNN-ResNet50Backbone=ResNet-50, Detector Architecture=Cascade R-CNN, Activation/Attention Module=APA*+AGLU2024.07 | 35.4 | — | — | — | |
| SE-CascadeRCNN-ResNet50Backbone=ResNet-50, Detector Architecture=Cascade R-CNN, Activation/Attention Module=SE2024.07 | 33.3 | — | — | — | |
| CascadeRCNN-ResNet50Backbone=ResNet-50, Detector Architecture=Cascade R-CNN2024.07 | 31.6 | — | — | — | |
| APA*+AGLU-FasterRCNN-ResNet50Backbone=ResNet-50, Detector Architecture=Faster R-CNN, Activation/Attention Module=APA*+AGLU2024.07 | 29.9 | — | — | — | |
| SE-FasterRCNN-ResNet50Backbone=ResNet-50, Detector Architecture=Faster R-CNN, Activation/Attention Module=SE2024.07 | 27 | — | — | — | |
| FasterRCNN-ResNet50Backbone=ResNet-50, Detector Architecture=Faster R-CNN2024.07 | 25.4 | — | — | — | |
| Cascade Mask R-CNNBackbone=EVA-VIT-G, Params=1.1B, D^backbone=Merged-30M, D^detector=Object3652024.12 | — | 49.4 | 54.8 | 51.4 | |
| DINO + ProvaBackbone=Swin-Base, Params=213.4M, D^backbone=ImageNet-22K2024.12 | — | 50.3 | 56.1 | 52.6 |