Mortal Kombat 3 Reinforcement Learning

Setup

1. Set up Python Environment

Create and activate a virtual environment:

python -m venv venv
source venv/bin/activate

2. Install DIAMBRA Dependencies

uv pip install diambra
uv pip install diambra-arena

3. Configure ROM Directory

Create the DIAMBRA roms directory:

mkdir -p /home/[USERNAME]/.diambra/roms/

You can use any other directory on your system as well, but then you need to set a system environment variable DIAMBRAROMSPATH=/absolute/path/to/roms/folder/

tip : For ROMs , find the game on diambra docs and check for search keywords and feed them to Gr0k , verify SHA ~ worked for mortal kombat 3 :)

Copy your UMK3 ROM file to the directory:

cp ~/path/to/umk3.zip /home/[USERNAME]/.diambra/roms/

4. Verify ROM Installation

Check if the ROM is properly installed:

diambra arena check-roms umk3.zip

If you are using a custom directory, and have set the environment varible correctly, this should run without errors as well. Try restarting the terminal if it doesn't.

✅ This command must pass successfully before proceeding. -_-

5. Register DIAMBRA Account

On first run, you'll need to register for a DIAMBRA account:

1. Run the test script: diambra run python trial.py (You need docker running)
2. When prompted to register, visit diambra.ai (cli gives diambra.ai/register it wont work)
3. Click "Login" (top right) and create an account
4. Complete the registration process

🎮 Usage

Run the RL training script:

diambra run -r /home/[USERNAME]/.diambra/roms/ python trial.py

or simply (if you have setup the environment variable)

diambra run python trial.py

Implemented A2C and PPO , you can find the code in the respective branches.

🔗 Resources

• DIAMBRA Arena Documentation

---