Skip to content
master
Switch branches/tags
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
 
 
 
 

A Gym environment for the new Arcade Learning Environment (v0.6.0)

A python Gym environment for the new Arcade Learning Environment (v0.6.0) supporting different difficulties and game modes. Enables experimenting with different Atari game dynamics within the Gym framework. This is fully inspired by the Atari environment in OpenAI gym.

Installation

  1. Install the latest Arcade Learning Environment by following the instructions at https://github.com/mgbellemare/Arcade-Learning-Environment.

  2. Download a ROM-file for the Atari game, or use a ROM file included in the repository. Title the ROM-file all lowercase, such as 'pong.bin'.

  3. Import UpdatedAtariEnv and follow the OpenAI gym API. An introduction to the new ALE is available here: https://arxiv.org/pdf/1709.06009.pdf.

This wrapper is initially developed and used in Sample-Efficient Deep RL with Generative Adversarial Tree Search.

The following is a list of the available modes and difficulties for each game from the new ALE paper.

ALE Modes and Difficulties

About

Environment for Atari on OpenAI gym using the new Arcade Learning Environment v0.6.0.

Resources

Releases

No releases published

Packages

No packages published

Languages