A Simple Single-Scale Vision Transformer for Object Localization and Instance Segmentation

Benchmark Model Rank Results
instance-segmentation-on-coco-minivalUViT-B+ + Cascade Mask R-CNN (w/ self-training)#56mask AP: 46.1
object-detection-on-coco-val2017UViT-B+ + Cascade Mask R-CNN (w/ self-training)#132box AP: 53.9
image-classification-on-imagenetUViT-B#690Top 1 Accuracy: 81.3%