Introduction to deep reinforcement learning using policy gradient methods.
The file evaluation.ipynb walks through implementation steps of the individual functions that are located in solution.py. The file walker_training.ipynb then serves as a template for the walker policy implementation, though it is nearly empty. You are allowed and expected to use any code from evaluation.ipynb to train the WalkerPolicy.
Only the files solution.py and WalkerPolicy.py should be uploaded into the evaluation system in a zip file. You can however also include the learned weights of your WalkerPolicy implementation and load them at initialization.
This homework has been developed using Python 3.10. You are advised to use it as well to avoid complications. Required libraries are listed in requirements.txt .