RT-DETRv4: Painlessly Furthering Real-Time Object Detection with Vision Foundation Models

Benchmark Model Rank Results
object-detection-on-coco-val2017RT-DETRv4-X#76box AP: 57.0
object-detection-on-coco-val2017RT-DETRv4-L#100box AP: 55.4
object-detection-on-coco-val2017RT-DETRv4-M#136box AP: 53.7
object-detection-on-coco-val2017RT-DETRv4-S#225box AP: 49.8
real-time-object-detection-on-coco-val2017RT-DETRv4-L–box AP: 55.4FPS (batch 1, GPU in brackets): 124 (T4)
real-time-object-detection-on-coco-val2017RT-DETRv4-M–box AP: 53.5FPS (batch 1, GPU in brackets): 169 (T4)
real-time-object-detection-on-coco-val2017RT-DETRv4-S–box AP: 49.7FPS (batch 1, GPU in brackets): 273 (T4)
real-time-object-detection-on-coco-val2017RT-DETRv4-X–box AP: 57.0FPS (batch 1, GPU in brackets): 78 (T4)