Break the Mold

Author	Jozef Porubcin

The goal of this project, Break the Mold, is to use reinforcement learning to train an AI model for close-quarters battle situations such as hostage rescue scenarios. I designed an environment in Unity to simulate such a scenario and I used the Unity ML-Agents Toolkit to train an AI model for this task. The environment supports two opposing teams of up to 12 agents each.

Controls

1. Move: WASD
2. Aim: Cursor position
3. Shoot: Primary mouse button
4. Reload: R

Training Setup

The Unity project can be opened in Unity 6 as-is. The following instructions are for setting up an environment for training an AI model with the Unity ML-Agents Toolkit (release 22).

1. Install Python 3.10.12. You can use pyenv to manage multiple Python versions:

   pyenv install 3.10.12
   

2. Set the local Python version:

   pyenv local 3.10.12
   

3. Create a virtual environment with Python 3.10.12:

   pyenv exec python -m venv .venv
   

4. Activate the virtual environment:

   Unix/macOS:

   source .venv/bin/activate
   

   Windows:

   .venv\Scripts\activate
   

5. Install mlagents 1.1.0:

   pip install mlagents==1.1.0
   

Now you can use the training configuration file to train an AI model with the Unity ML-Agents Toolkit. Replace the default model file to test a different model.