How to Train a Custom Faster RCNN Model In PyTorch

Author(s): Dr. Leon Eversberg

Originally published on Towards AI.

Fine-tuning a pre-trained Faster RCNN model with custom images in the COCO data format using PyTorch
Training and validation loss during model training. Source: Author

In this PyTorch tutorial for beginners, we will use a pre-trained object detection model from Torchvision and fine-tune it on a custom image dataset in the COCO data format.

First of all, is it worth learning PyTorch, or should you learn another framework like TensorFlow?

According to Papers With Code, PyTorch is currently the leading deep learning framework for academic paper implementations. In fact, PyTorch is by far the leading framework with 61%. Only 4% of papers are currently implemented in TensorFlow.

Additionally, PyTorch has overtaken TensorFlow in the global Google Trends statistics in 2021.

If you want to learn deep learning right now, PyTorch is the way to go.

To train an object detection model, we first need a dataset containing images and bounding box annotations.

One of the most commonly used dataset formats is the Microsoft COCO benchmark dataset [1].

The official COCO dataset format for object detection annotations is as follows:

image{ "id": int, "width": int, "height": int, "file_name": str, "license": int, "flickr_url": str, "coco_url": str, "date_captured": datetime,}annotation{ "id": int, "image_id": int, "category_id": int, "segmentation": RLE or [polygon], "area": float, "bbox": [x,y,width,height], "iscrowd": 0 or 1,}categories[{ "id": int, "name": str, "supercategory": str,}]

Each image requires an id and… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI