| object-detection-on-coco-o | DyHead
(Swin-L) | #9 | Average mAP: 35.3Effective Robustness: 10.00 |
| object-detection-on-coco-o | DyHead
(ResNet-50) | #31 | Average mAP: 19.3Effective Robustness: 0.16 |
| object-detection-on-coco-test-dev | DyHead (Swin-L, multi scale, self-training) | #24 | box mAP: 60.6AP50: 78.5AP75: 66.6APM: 64.0APL: 74.2 |
| object-detection-on-coco-test-dev | DyHead (Swin-L, multi scale) | #32 | box mAP: 58.7AP50: 77.1AP75: 64.5APM: 62.0APL: 72.8 |
| object-detection-on-coco-test-dev | DyHead (ResNeXt-64x4d-101-DCN, multi scale) | #56 | box mAP: 54AP50: 72.1AP75: 59.3 |
| object-detection-on-coco-test-dev | DyHead (ResNeXt-64x4d-101) | #106 | box mAP: 47.7AP50: 65.7AP75: 51.9 |
| object-detection-on-coco-test-dev | DyHead (ResNet-101, 2x) | #116 | box mAP: 46.5AP50: 64.5AP75: 50.7APS: 28.3APM: 50.3APL: 57.5 |
| object-detection-on-coco-test-dev | DyHead (ResNet-50) | #148 | box mAP: 43AP50: 60.7AP75: 46.8 |
| object-detection-on-coco-val2017 | DyHead (Swin-L, multi scale, self-training) | #28 | box AP: 60.3 |
| object-detection-on-coco-val2017 | DyHead (Swin-L, multi scale) | #52 | box AP: 58.4 |
| object-detection-on-coco-val2017 | DyHead (Swin-T, 2x) | #226 | box AP: 49.7AP50: 68.0AP75: 54.3APS: 33.3APM: 54.2APL: 64.2 |