Reinforcement Learning with Python, scikit-learn, Godot, and C#

Project Overview

This GitHub repository is dedicated to the integration of Reinforcement Learning (RL) using Python and scikit-learn with the Godot game engine using C#. This project aims to demonstrate how to leverage RL techniques to create intelligent agents within the Godot game environment. By combining the powerful RL capabilities of scikit-learn with the game development potential of Godot, you can create interactive and adaptive game experiences.

The project uses vanilla crossentropy(CE) method for policy approximated by a neural network. For RL, it boild down to Repeat:

• Generate N games
• Take M best
• Fit to those M best samples

Table of Contents

• Getting Started
  • Prerequisites
  • Installation
• Usage
• Contributing
• License

Getting Started

Prerequisites

Before you begin, ensure you have met the following requirements:

• Python 3.x installed on your system.
• scikit-learn library installed.
• Godot Engine installed on your machine.
• Basic knowledge of Python, scikit-learn, C#, and the Godot engine.

Installation

1. Clone this GitHub repository to your local machine:

   git clone [email protected]:HectorPulido/godot_rl.git

2. Open the rl_client folder in the Godot Engine editor.

3. Explore.

Usage

1. Setting up the Python Environment: First, set up your Python environment:

cd rl_server
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
python3 main.py

You can touch the code in Bot class to change the algorithm.

2. Integrating with Godot: In the rl_client folder, you'll find example Godot projects and C# scripts that demonstrate how to interface with the Python RL agents. You can use these examples as a starting point for your own Godot projects.

3. Testing and Iterating: Experiment with your RL agents within the Godot environment. Tweak the integration code and the RL algorithms to achieve the desired behavior for your game.

Contributing

Contributions to this project are welcome! If you have any ideas, bug fixes, or improvements, please open an issue or submit a pull request.

License

This project is licensed under the MIT License - see the LICENSE file for details.

---

Let's connect 😋