UnityRL
My reinforcement learning experiments with Unity ML Agents Toolkit. Training reinforcement learning agents to land a rocket in 3D environment!
https://github.com/Unity-Technologies/ml-agents
Usage
Open Project folder in Unity
After project is open, press play button to see trained agent in action.
Training
First follow these install instructions: https://github.com/Unity-Technologies/ml-agents/blob/release_12_docs/docs/Installation.md
Now you can run
mlagents-learn config/lunarlander_config.yaml --run-id=lunarlander
and after it says it, press play button in Unity
After training model is saved in a folder called "results".
You can set the model in Unity for LunarLander prefab. In the prefab under "Behavior Parameters" drag the new model to "model".
For better advice read here: https://github.com/Unity-Technologies/ml-agents/blob/release_12_docs/docs/Getting-Started.md