This repository contains a simple game, wrapped to match the OpenAI Gym API.
The player must fill all vacant cells of the board. He must avoid trying to fill a non-vacant cell. An empty cell is represented by "0" and a filled cell by "1". In the example below, only the second cell is filled, and so the agent must select any of the other cells:
from gym_game import Game
env = Game()
env.reset()
env now functions similar to any basic gym environment.
git clone https://github.com/shakedzy/gym_game.git
pip install ./gym_game
Dependencies: numpy