RLGymExampleBot is an example bot for Rocket League, built using the RLGym library. This project provides a framework for developing, training, and deploying reinforcement learning agents for Rocket League. It's designed to be a starting point for researchers and enthusiasts looking to experiment with RL agents in the game.

• RLGym Integration: Seamlessly integrates with the RLGym library for environment interaction.
• Configurable Agents: Easily switch between different observation builders, action parsers, and models via configuration.
• Model Loading: Supports loading the latest trained model or a specific model file.
• Training Configurations: Includes example configurations for various training algorithms (e.g., PPO, SAC, TD3).
• RLBot Integration: Designed to run within the RLBot framework.

To set up the RLGymExampleBot, follow these steps:

1. Clone the repository:

   git clone https://github.com/your-repo/RLGymExampleBot.git
   cd RLGymExampleBot

2. Install dependencies: The run.py script will automatically check for and install the necessary Python dependencies listed in requirements.txt when you first run the bot.

To run the RLGymExampleBot, execute the run.py script:

python run.py

This will launch the RLBot framework and load the RLGymExampleBot. Ensure that Rocket League is running and the RLBot framework is properly set up.

This project supports training reinforcement learning models using various algorithms. Training configurations are located in the configs/ directory. You can launch training using the src/launch_training.py script (or similar, depending on your setup).

Example training command (assuming src/launch_training.py is the entry point):

python src/launch_training.py --config configs/standard_training.json

Refer to the configs/ directory for available training configurations and src/train.py for the training script details.

The main configuration for the bot is located in src/bot.cfg. This file allows you to customize:

• model_path: The path to your trained model. Set to latest to automatically load the most recently created model from the models directory.
• model_dir: The directory where your trained models are stored (defaults to ../models).
• tick_skip: The number of game ticks to skip between bot decisions.
• obs_builder: The observation builder class to use (e.g., DefaultObs, AdvancedObs).
• act_parser: The action parser class to use (e.g., DefaultAction, ContinuousAction).

RLGymExampleBot/
├── run.py                  # Main script to launch the bot
├── requirements.txt        # Python dependencies
├── src/
│   ├── bot.py              # Main bot class
│   ├── rlbot_support.py    # Core RLBot agent logic, including model loading
│   ├── bot.cfg             # Bot configuration file
│   ├── train.py            # Script for training RL models
│   ├── launch_training.py  # Script to launch training with specific configs
│   ├── action/             # Action parsers
│   ├── obs/                # Observation builders
│   └── ...
├── configs/                # Training configuration files
├── models/                 # Directory for trained models
├── league/                 # (Potentially) League management system
├── gui.py                  # (Potentially) Graphical User Interface
├── docs/                   # Project documentation
└── ...