How to Train YOLO v5 on a Custom Dataset | DigitalOcean

Oct. 8, 2024 update: This tutorial now features some deprecated code for sourcing the dataset. Please see our updated tutorial on YOLOv7 for additional instructions on getting the dataset in a Jupyter Notebook for this demo.

Introduction

YOLO, or You Only Look Once, is one of the most widely used deep learning-based object detection algorithms. In this tutorial, we will go over how to train one of its latest variants, YOLOv5, on a custom dataset. More precisely, we will train the YOLO v5 detector on a road sign dataset. By the end of this post, you shall have an object detector that can localize and classify road signs. Before we begin, let me acknowledge that YOLOv5 attracted a lot of controversy when it was released over whether it’s right to call it v5. I’ve addressed this a bit at the end of this article. For now, I’d simply say that I’m referring to the algorithm as YOLOv5 since that is the name of the code repository.