Reinforcement learning without the install headache
It exists because learning reinforcement learning should not start with dependency hell.

What it does This repository houses five Google Colab notebooks originally built for the 2019 Journées Nationales de la Recherche en Robotique. Each notebook covers a specific Stable-Baselines3 topic—getting started, Gym wrappers, saving and loading models, multiprocessing, callbacks and hyperparameter tuning, and custom environments—so you can run code directly in a browser. A sixth bonus notebook explores the RL Baselines3 Zoo.
The interesting bit The project treats a conference tutorial as living documentation. Because everything runs in Colab, the gap between reading about RL and actually breaking an environment shrinks to a single click.
Key highlights
- Five progressive notebooks covering fundamentals through custom environments
- Browser-first execution via Google Colab; no local install required
- Accompanied by slide deck and links to upstream Stable-Baselines3 docs
- Bonus notebook on RL Baselines3 Zoo for pre-trained model exploration
- Originally presented at JNRR 2019 by core Stable-Baselines contributors
Verdict Reach for this if you want to experiment with Stable-Baselines3 before committing to a local setup. If you already know the API and need deep reference material, head straight to the official documentation instead.
Frequently asked
- What is araffin/rl-tutorial-jnrr19?
- It exists because learning reinforcement learning should not start with dependency hell.
- Is rl-tutorial-jnrr19 open source?
- Yes — araffin/rl-tutorial-jnrr19 is open source, released under the MIT license.
- What language is rl-tutorial-jnrr19 written in?
- araffin/rl-tutorial-jnrr19 is primarily written in Jupyter Notebook.
- How popular is rl-tutorial-jnrr19?
- araffin/rl-tutorial-jnrr19 has 747 stars on GitHub.
- Where can I find rl-tutorial-jnrr19?
- araffin/rl-tutorial-jnrr19 is on GitHub at https://github.com/araffin/rl-tutorial-jnrr19.