acrobot

Use of Deep Q-Learning to build a RL Agent. Case of the Gym Acrobot-v1 environment.

Installation

Install the project environment

conda env create -f env.yml
conda activate acrobot

Training

Train the agent with

python train.py

The agent pytorch model will be saved in a file called state.pt. You can change this name and other parameter in the file train.py.

Simulation

Simulate the environment and the agent with

python test.py