Focal Self-attention for Local-Global Interactions in Vision Transformers

Benchmark Model Rank Results
instance-segmentation-on-coco-test-devFocal-L (HTC++, multi-scale)#18mask AP: 51.3AP50: 75.4AP75: 56.5APS: 35.6APL: 64.2
instance-segmentation-on-coco-minivalFocal-L (HTC++, multi-scale)#20mask AP: 50.9
semantic-segmentation-on-ade20k-valFocal-L (UperNet, ImageNet-22k pretrain)#29mIoU: 55.4
object-detection-on-coco-test-devFocal-L (DyHead, multi-scale)#30box mAP: 58.9
object-detection-on-coco-val2017Focal-L (DyHead, multi-scale)#44box AP: 58.7
semantic-segmentation-on-ade20kFocal-L (UperNet, ImageNet-22k pretrain)#48Validation mIoU: 55.40