This library is designed for education purposes, it is mainly used to perform some practical experiences with various RL algorithms. It facilitates using optuna for tuning hyper-parameters and using rliable and statistical tests for analyzing the results.
git clone https://github.com/osigaud/bbrl_algos.git
cd bbrl_algos
pip install -e .
We suggest using your favorite python environment (conda, venv, ...) as some further installations might be necessary
go to src/bbrl_algos, choose your algorithm and run python3 your_algorithm.py