README.md

SARSA to beat gravity 🚃

👉 Read in datamachines 👉 Read in Towards Data Science

This is part 2 of my course Hands-on reinforcement learning.

In this part we use SARSA to help a poor car win the battle against gravity!

Be like a train; go in the rain, go in the sun, go in the storm, go in the dark tunnels! Be like a train; concentrate on your road and go with no hesitation!

--Mehmet Murat Ildan

Quick setup

The easiest way to get the code working in your machine is by using Poetry.

1. You can install Poetry with this one-liner:

   $ curl -sSL https://raw.githubusercontent.com/python-poetry/poetry/master/get-poetry.py | python -

2. Git clone the code

   $ git clone https://github.com/Paulescu/hands-on-rl.git 

3. Navigate to this lesson code 02_mountain_car

   $ cd hands-on-rl/02_mountain_car

4. Install all dependencies from `pyproject.toml:

   $ poetry install

5. Activate the virtual environment

   $ poetry shell

6. Set PYTHONPATH and launch jupyter (jupyter-lab param may fix launch problems on some systems)

   $ export PYTHONPATH=".."
   $ jupyter-lab --NotebookApp.use_redirect_file=False

Notebooks

1. Explore the environment
2. Random agent baseline
3. SARSA agent
4. Momentum agent
5. Homework