Skip to content

Latest commit

 

History

History
39 lines (22 loc) · 1.17 KB

ReadMe.md

File metadata and controls

39 lines (22 loc) · 1.17 KB
Raw

Unity Project: Reinforcement Learning on Circuit de Barcelona-Catalunya

This project showcases a car agent that has been trained using reinforcement learning to complete successful laps on a scaled-down version of the Circuit de Barcelona-Catalunya. We've used the mlagents package from Unity for the agent's training.

Trained Lap Simulation

Requirements:

  • Unity (2022.3.9f1)
  • mlagents package installed

Getting Started:

1. Clone the Repository:

Ensure that you've cloned this repository to your local machine.

2. Navigate to the Build Folder:

Open a terminal or command prompt and navigate to the Build folder of the cloned project.

3. Train the Model:

To train the model yourself, run the following command:

mlagents-learn config.yaml --run-id=<YOUR_DESIRED_NAME>

Experience the Trained Model:

A pre-trained model is included in the project. To see the car agent in action:

  1. Open the project in Unity.
  2. Play the environment from the Unity interface.

Enjoy the simulation and feel free to contribute or suggest improvements!