AWS DeepRacer
Developer Guide

This is prerelease documentation for a service in preview release. It is subject to change.

Train and Evaluate AWS DeepRacer Models Using Amazon SageMaker Notebooks

The AWS DeepRacer console provides you with an integrated experience to train and evaluate your AWS DeepRacer models. It's integrated because AWS DeepRacer uses Amazon SageMaker and AWS RoboMaker behind the scenes. The integration includes detailed reinforcement learning tasks and makes training more readily accessible to beginners.

If you're an experienced user of Amazon SageMaker or if you're determined to learn how to use Amazon SageMaker and AWS RoboMaker to train and evaluate your AWS DeepRacer models, then you can manually create an Amazon SageMaker notebook. You can then clone a reinforcement learning sample notebook instance and use it as a template to perform the predefined tasks that train and evaluate an AWS DeepRacer model.

After the training, you can copy the trained model artifacts to your AWS DeepRacer vehicle for test runs in a physical environment.

The tutorial presents step-by-step instructions to walk you through these tasks.