Object detection in an urban environment

In this project, you will learn how to train an object detection model using the Tensorflow Object Detection API and AWS Sagemaker. At the end of this project, you will be able to generate videos such as the one below.

Installation

Refer to the Setup Instructions page in the classroom to setup the Sagemaker Notebook instance required for this project.

Note: The conda_tensorflow2_p310 kernel contains most of the required packages for this project. The notebooks contain lines for manual installation when required.

Usage

This repository contains two notebooks:

• 1_train_model: this notebook is used to launch a training job and create tensorboard visualizations.
• 2_deploy_model: this notebook is used to deploy your model, run inference on test data and create a gif similar to the one above.

First, run 1_train_model.ipynb to train your model. Once the training job is complete, run 2_deploy_model.ipynb to deploy your model and generate the animation.

Each notebook contains the instructions for running the code, as well the requirements for the writeup.

Note: Only the first notebook requires a write up.

Useful links

• The Tensorflow Object Detection API tutorial is a great resource to debug your code. This section in particular will teach you how to edit the pipeline.config file to update your training job.

• This blog post teaches how to label data, train and deploy a model with the Tensorflow Object Detection API and AWS Sagemaker.