Transfer-Learning-Atari-Games

It can emulate any of the following games:

['Asterix', 'Asteroids', 'MsPacman', 'Kaboom', 'BankHeist', 'Kangaroo', 'Skiing', 'FishingDerby', 'Krull', 'Berzerk', 'Tutankham', 'Zaxxon', 'Venture', 'Riverraid', 'Centipede', 'Adventure', 'BeamRider', 'CrazyClimber', 'TimePilot', 'Carnival', 'Tennis', 'Seaquest', 'Bowling', 'SpaceInvaders', 'Freeway', 'YarsRevenge', 'RoadRunner', 'JourneyEscape', 'WizardOfWor', 'Gopher', 'Breakout', 'StarGunner', 'Atlantis', 'DoubleDunk', 'Hero', 'BattleZone', 'Solaris', 'UpNDown', 'Frostbite', 'KungFuMaster', 'Pooyan', 'Pitfall', 'MontezumaRevenge', 'PrivateEye', 'AirRaid', 'Amidar', 'Robotank', 'DemonAttack', 'Defender', 'NameThisGame', 'Phoenix', 'Gravitar', 'ElevatorAction', 'Pong', 'VideoPinball', 'IceHockey', 'Boxing', 'Assault', 'Alien', 'Qbert', 'Enduro', 'ChopperCommand', 'Jamesbond']

Objective

The objective of this project is to implement and compare various RL approaches with atari games as a common environment, and to measure the impact of transfer learning across similar objective games.

Demo Video

https://youtu.be/EOIF9reE8zI

Usage

1. Clone the repo.
2. Go to the project's root folder.
3. Install required packages pip install -r requirements.txt.
4. Training example - python3 atari.py -g=Pong -m=ddqn_train.
5. Testing example - python3 atari.py -g=Pong -m=ddqn_test -r=True.
6. Help python atari.py --help.

Model Architecture

Deep Convolutional Neural Network by DeepMind

Know issues

1. When last tried Tensorflow didn't support 3.7.x, use 3.6 versions