The dataset consists of 513 images, each with a text file describing the image’s label. In the text file, each row describes a
bounding-box in Yolo’s format.
• One row per object
• Each row is class xcenter ycenter width height format
• Box coordinates must be in normalized xywh format (from 0 - 1)
• Class numbers are zero-indexed (start from 0)
There are a total of more than 500 images in the dataset. We have randomly split them into train, validate and test sets. So
they can be used to train the Yolov5 model immediately