Skip to main content

OpenAI Gym Environment for 2048

Project description

https://travis-ci.com/activatedgeek/gym-2048.svg?branch=master https://badge.fury.io/py/gym-2048.svg

This package implements the classic grid game 2048 for OpenAI gym environment.

Install

pip install gym-2048

Environment(s)

The package currently contains two environments

  • Tiny2048-v0: A 2 x 2 grid game.

  • 2048-v0: The standard 4 x 4 grid game.

Attributes

  • Observation: All observations are n x n numpy arrays representing the grid. The array is 0 for empty locations and numbered 2, 4, 8, ... wherever the tiles are placed.

  • Actions: There are four actions defined by integers.
    • LEFT = 0

    • UP = 1

    • RIGHT = 2

    • DOWN = 3

  • Reward: Reward is the total score obtained by merging any potential tiles for a given action. Score obtained by merging two tiles is simply the sum of values of those two tiles.

Rendering

Currently only basic print rendering (mode='human') is supported. Graphic rendering support is coming soon.

Usage

Here is a sample rollout of the game which follows the same API as OpenAI gym.Env.

import gym_2048
import gym


if __name__ == '__main__':
  env = gym.make('2048-v0')
  env.seed(42)

  env.reset()
  env.render()

  done = False
  moves = 0
  while not done:
    action = env.np_random.choice(range(4), 1).item()
    next_state, reward, done, info = env.step(action)
    moves += 1

    print('Next Action: "{}"\n\nReward: {}'.format(
      gym_2048.Base2048Env.ACTION_STRING[action], reward))
    env.render()

  print('\nTotal Moves: {}'.format(moves))

NOTE: Top level import gym_2048 is needed to ensure registration with Gym.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

gym-2048-0.2.6.tar.gz (4.6 kB view details)

Uploaded Source

File details

Details for the file gym-2048-0.2.6.tar.gz.

File metadata

  • Download URL: gym-2048-0.2.6.tar.gz
  • Upload date:
  • Size: 4.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/1.13.0 pkginfo/1.5.0.1 requests/2.21.0 setuptools/40.8.0 requests-toolbelt/0.9.1 tqdm/4.31.1 CPython/3.6.3

File hashes

Hashes for gym-2048-0.2.6.tar.gz
Algorithm Hash digest
SHA256 13b15d0cd5ee556df33eab6861bfa94ae88aeb063fc1fd6d3e7fee230a0fa662
MD5 75c53e794d6d77db06bb90009b597b2a
BLAKE2b-256 4747a1f06f3a6a05b78ffee842a7cd6f6734f97b0c3600f69df7654c232f0adc

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page