This tutorial is based on the YOLOv7 repository by WongKinYiu. This notebook shows training on your own custom objects.
In this tutorial, it will be utilized an open source computer vision dataset available on Roboflow Universe in the website https://universe.roboflow.com/university-bswxt/crack-bphdr, which contains more than 4000 images, train set, valid set, and test set, with the labels for the segmentation.
Crack Segmentation with Yolo V8, best weight = https://drive.google.com/file/d/1XWHCK-MTiz2AJqJyz5LT-sAFLeFef91E/view?usp=sharing