Skip to content

A simple OpenAI Gym environment for single and multi-agent reinforcement learning

License

Notifications You must be signed in to change notification settings

kevin0748/slimevolleygym

 
 

Repository files navigation

Slime Volleyball Gym Environment

Installation

The program runs on Python 3.7. Install all dependency to do training and evaluation:

git clone https://github.com/JStrick510/slimevolleygym.git
cd slimevolleygym
pip install -e .

Basic Usage

Train the agents:

# Train agent against PPO
python training_scripts/train_w_best_opponent.py

# Train agent against GA
python training_scripts/train_w_best_ga.py

Evaluation the agents:

# Evaluate ppo_new against baseline
python eval_agents.py --left baseline --right ppo_new

# Evaluate ppo_ga against ppo_new
python eval_agents.py --left ppo_new --right ppo_ga 

# Evaluate ppo_ga against ppo_new with render
python eval_agents.py --left ppo_new --right ppo_ga --render

Training Result

Training result are located under folder best_opponent_vs_ppo1 and best_opponent_ga:

  1. final_model.zip: the final model to evaluate in the game
  2. progress.csv: the progress of training

Folder Structure:

├── best_opponent_ga
│   ├── final_model.zip
│   └── progress.csv
├── best_opponent_vs_ppo1
│   ├── final_model.zip
│   └── progress.csv
├── setup.py
├── training_scripts
└── zoo

About

A simple OpenAI Gym environment for single and multi-agent reinforcement learning

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%