Submitting Your Agent Team

This page gives you an overview on how to format your entry for submission to our competition site. If you need information on training your entry please see Training Your Entry page.

Understanding the solution.py file

rl_test/solution.py

Solution Initialization: __init__; should load in any necessary heuristics or neural networks to be used when compute action is called

compute action: full observation normalized and unormalized, the last argument is the full game state with absolute positions if you use this ensure your solution works on both RED and BLUE sides

Formatting Your Submission

Example files for training your agent can be seen in

rl_test/competition_info/example_submission.zip

Inside each zip file should be the follwing pieces:

• solution.py
• config.py
• Any Neural Network Or Supporting .py files

 1.) Update solution.py to work with your approach

 2.) Select and zip your solution.py and the policy checkpoints your submission uses. DO NOT USE tar, gzip, or any other compressed folder format.

 3.) Submit your zip file to the competition site!