https://github.com/steven-hewitt/multi-agent-pong-rally
A multi-agent Gym-compatible environment based off of Atari's Pong.
https://github.com/steven-hewitt/multi-agent-pong-rally
Last synced: over 1 year ago
JSON representation
A multi-agent Gym-compatible environment based off of Atari's Pong.
- Host: GitHub
- URL: https://github.com/steven-hewitt/multi-agent-pong-rally
- Owner: Steven-Hewitt
- License: gpl-3.0
- Created: 2017-11-04T20:56:26.000Z (over 8 years ago)
- Default Branch: master
- Last Pushed: 2017-12-03T20:09:39.000Z (over 8 years ago)
- Last Synced: 2025-01-29T02:58:44.082Z (over 1 year ago)
- Language: Python
- Size: 37.1 KB
- Stars: 3
- Watchers: 2
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Multi-Agent-Pong-Rally (MAPR)
### A Gym environment.
MAPR is a Gym environment designed to emphasize collaboration between two actors.
Each actor has one action dimension, which represents an acceleration value for their respective paddle on the board.
## Installation Instructions
Each of the files in this repository (except `test.py`, which can be placed anywhere) has a header line detailing where in your Gym installation folder they should be placed. **`envs__init__.py` and `scoreboard__init__.py` both assume that you do not have any other custom gym environments installed! If you have other custom gym environments, follow gym custom environment installation instructions found [here](https://github.com/openai/gym/wiki/Environments).**
MAPR is not currently compatible with `gym_pull`, but I welcome pull requests that seek to resolve that issue.
Thanks to https://gist.github.com/vinothpandian/4337527 for the base code used for this gym environment.