Skip to content

Latest commit

 

History

History
37 lines (24 loc) · 685 Bytes

README.md

File metadata and controls

37 lines (24 loc) · 685 Bytes

Refinforcement Learning to play Atari 2600 games

Dependencies

  • OpenAI Gym (need Pong, Breakout and Space Innvaders ROMs)
  • Python3 virtualenv (Recommend using pipenv)

Install

First, clone the reposiory and run the following command inside the directory

$ pipenv --python 3.7
$ pipenv install
$ pipenv lock --pre

Launch Visualisation

$ python3 app.py

Begin training

Train DQN:

$ python3 agents/runner.py --algo=dqn

Train DQN with Double Q-Learning:

$ python3 agents/runner.py --algo=dqn --doubledqn

Train Duelling DQN:

$ python3 agents/runner.py --algo=ddqn

Run the command: $ python3 agents/runner.py --help for more information