PyTorch Object Detection :: YOLOv5 TXT

YOLOR

What is YOLOR?

You Only Learn One Representation (YOLOR) is a state-of-the-art object detection model. YOLOR pre-trains an implicit knowledge network with all of the tasks present in the COCO dataset, namely object detection, instance segmentation, panoptic segmentation, keypoint detection, stuff segmentation, image caption, multi-label image classification, and long-tail object recognition. When optimizing for the COCO dataset, YOLOR trains another set of parameters that represent explicit knowledge. For prediction, both implicit and explicit knowledge is used.

YOLOR Architecture

YOLOR Architecture

Vision Transformer Performance

This novel approach propels YOLOR to the state-of-the-art for object detection in the speed/accuracy tradeoff landscape.
YOLOR Performance
Images in Courtesy of Wong-Kin-Yiu

Further Reading

Train YOLOR on a Custom Dataset: https://blog.roboflow.com/train-yolor-on-a-custom-dataset/
YOLOR Research Paper: https://arxiv.org/abs/2105.04206