I used to watch a lot of code bullet videos, where he created evolutionary machine learning models to be the best a simple games. When I heard this project essentially wanted us to do this but for Tron, I decided to use a similar approach, using pyTorch and gymnasium to train my agent against itself to be the best it can be. However using these software were a lot more difficult to set up than I anticipated, and the training ended up taking far longer than just one night with the method I used. I ended up submitting a sample bot that I used to train my bot against, which is unfortunate. However I learned a lot! I now know the basics of software I will probably use in my future career, and the basics of machine learning.
Case-closed
A bot that plays the game tron to the best of it's ability based, on data collected from millions of learning runs.
Updates
Leave feedback in the comments!
Log in or sign up for Devpost to join the conversation.