Do you like sumo ?

Do you like wonky physics and hover cars!?

Do you like Reinforcement Learning!?

Then you'll love stupid car sumo simulation trained with Unity MLAgents and Proximal Policy Optimization

This is an experiment I bodged together where I tried to train two competitive agents using the same Neural Network with a lot of trial and error to try and get a working policy.