Parkour Beans?

This is a machine learning project created in Unity. Basically, there are a bunch of non-player characters, and they figure out how to do parkour. And you have complete control over their training. Right now the parkour is simple so that it does take very long to train, but I plan to add some more levels soon. Also, when you start the program there will be a bunch of buttons. Population size refers to the number of NPCs. The more, the quicker it will train, but will also be more computationally taxing. Mutation rate refers to how much randomness is in your program. The more randomness, the more certain it will learn, but it will take longer. Lifespan is how long the NPCs live each generation in frames. Speed is how fast they can go. And jump height is how high they can jump. As you can probably tell, its really about finding a balance. I have it preset on settings that should work on most computers, and trains pretty quickly. But if you have a beefy computer, always feel free to crank up that population size. It uses a genetic algorithm to train the NPCs. If you curious about that, watch this video. (You might want some previous knowledge of programming).