Real-Time Object Detection Meets DINOv3

Benchmark Model Rank Results
real-time-object-detection-on-coco-val2017DEIMv2-X (DINOv3 backbone)#7box AP: 57.8
real-time-object-detection-on-coco-val2017DEIMv2-L (DINOv3 backbone)#17box AP: 56.0
real-time-object-detection-on-coco-val2017DEIMv2-M (DINOv3-distilled backbone)#52box AP: 53.0
real-time-object-detection-on-coco-val2017DEIMv2-S (DINOv3-distilled backbone)#80box AP: 50.9
real-time-object-detection-on-coco-val2017DEIMv2-Nano#123box AP: 43.0
real-time-object-detection-on-coco-val2017DEIMv2-Pico#144box AP: 38.5
object-detection-on-coco-val2017DEIMv2-X#63box AP: 57.8
object-detection-on-coco-val2017DEIMv2-L#90box AP: 56.0
object-detection-on-coco-val2017DEIMv2-M#156box AP: 53.0
object-detection-on-coco-val2017DEIMv2-S#200box AP: 50.9