I have a question about yolo head also for all object detection. when reading yolov4 source,in detection head,using many "reshape" operaction and get shape like batch_
