Object Detection Using Faster RCNN in Pytorch

Introduction

This is a tutorial for beginners on how to train a Faster RCNN network for object detection in pytorch. Dataset used is Global Wheat Detection available on kaggle. The code here is commented with explanations and customised visualizing functions for validation image output after every epoch.

Note: Incase notebook is not loading on GitHub, you can check notebook with validation output upto 10 epochs here.

Code in python file has some issues so check notebook for smoothly working code.

Please STAR ⭐ the repository.

Requirements

• Python(3.6+)
• Pytorch(1.7)
• GPU: Nvidia Tesla P100(provided by Kaggle)

Try Yourself

Download the saved model from here.

Validation Output Sample

Test Output Sample