This project demonstrates how to use the new Experimental AnimGen plugin which can be used to build and train Machine-Learning-based character animation controllers in the editor and evaluate them at runtime.

It contains two examples of AnimGen animation controllers:

• A player controller that supports basic walking locomotion as well as aiming and firing of a pistol. This controller can be re-trained within the sample, as all training data is provided.

• A pre-trained NPC controller that supports walking locomotion in a very large number of different styles including support for props. This controller cannot be re-trained in the editor as training data is not provided, but acts as an example of the scalability of the system, and a template for a more complex setup.