No description, website, or topics provided.
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
gym_magic
LICENSE
README.md
play-magic-square.py
setup.py
train-magic-square-DQN.py

README.md

This repository contains a PIP package which is an OpenAI environment for simulating an environment in which a magic-square is solved.

Installation

Install the OpenAI gym.

Then install this package via

pip install -e .

REQUIRED LIBRARIES

  • numpy
  • keras
  • tensorflow
  • gym: pip install gym
  • tabulate: pip install tabulate

Usage

import gym
import gym_magic

env = gym.make('MagicSquare3x3-v0')

Train

python train-magic-square-DQN.py

Test

  1. Play the game
python play-magic-square.py

The Environment

...

CREDITS

Part of the code for Deep-Q-Learning is based on this blog-post.

REFERENCES

  1. "Human-level control through deep reinforcement learning", Nature 518 (7540), 529-533
  2. "Playing Atari with Deep Reinforcement Learning", preprint, arXiv:1312.5602
  3. Deep Reinforcement Learning Blog

TODO

  1. Tinker Brain's architecture
  2. Implement graphics window for play-magic-sqaure.py
  3. For further improvement ideas check out these: