Implementation and performance evaluation of Deep Q-Learning and Double DQN algorithms in LunarLander-v2 environment.
In order to run the notebooks, you can use the latest official TF-Docker container, with both Tensorflow 2.4.1 and Jupyter Notebook.
Python packages:
gym
gym-notebook-wrapper
matplotlib
numpy
keras
Box2D
PyVirtualDisplay
Ubuntu .deb
package (for a Docker-based env. on Windows):
xvfb