Object Detection with RT-DETR on COCO

An investigation into RT-DETR (Real-Time Detection Transformer), a fully end-to-end transformer architecture for object detection. Evaluated on the COCO dataset with a focus on understanding the trade-offs between detection accuracy and inference speed under hardware constraints (RTX 3050, 16 GB RAM).

Key topics: transformer-based detection · RT-DETR · COCO benchmark · speed-accuracy trade-off

Tech stack: Python · PyTorch · COCO · RT-DETR