PyTorch Object Detection :: Scaled-YOLOv4
Scaled YOLOv4 consists of a series of object detection models of varying size that are the new state of the art in object detection.
Scaled YOLOv4 is an extension of the YOLOv4 research, implemented in the YOLOv5 PyTorch framework. The primary research contribution is to implement Cross Stage Partial networks in the YOLOv4 architecture.
More info here: https://blog.roboflow.com/scaled-yolov4-tops-efficientdet/