Have you ever played Run 3 on Coolmath Games? Well so have we, but we are pretty bad. So we're building a Reinforcement Learning-based Agent to play Run's "Infinite Mode" for us.

Our agent will learn off of raw visual information, and will only take in a screenshot of the game every few frames. We anticipate this to be challenging as Run 3 is a 3d parkour game, and thus visual information will be tough to learn from.

Dw were locked in tho.

Built With

Share this project:

Updates