Model training

Unicorn Games has a tight deadline to justify a round of funding. Your team needs to show a functional demo of the game to investors. Because of this, you need to shorten the training time to 16 minutes and change the hyperparameters and the weights to get the best RL agent within an 16 minute time frame. Ideally, you would have more time to train, but you need to show your investors in three hours. This is a constraint that you must work within for now.

Once you are finished tuning your hyperparameters, continue on in the SageMaker notebook to build the Docker container image that your RL model will exist in and train the RL model. Detailed, step-by-step instructions are in the notebook.

Detailed, step-by-step instructions are in the SageMaker notebook called lunarlander-tutorial.ipynb