This project was merged from a existed PFNN agent project in Unity. To run the Scripts you need install Unity and Unity ML-agents first.

To run PFNN, you need click the store parameter buttoms on Adam's inspector window.

To run ML-Agents, please follow the instruction on： https://github.com/Unity-Technologies/ml-agents/blob/main/docs/ML-Agents-Overview.md

The main developments in Unity included Map, obstacles, walls .... can be view in Unity client, SIGGRAPH_2017/Unity/Asstes/Demo/Adam/Demo_Adam

To view the main developments of Reinforcement learning, please view SIGGRAPH_2017/Unity/Asstes/Scripst/Control

To view the results of training, open tensorboard in, SIGGRAPH_2017/Unity/results

To view the Demostration record for imitation learning, SIGGRAPH_2017/Unity/Asstes/Demos

Folder with different name of number indicated the Observation space size of the recorded demo.

To view the Hyper-parameters for training, check SIGGRAPH_2017/Unity/config