mayank15147 / save-me-ai

Implementation of Deep Q-Learning to Learn how to play a simple game written in python.

Save-me-AI

Implementation of Deep Q-Learning to Learn how to play a simple game written in python.

Here is the demo computer playing the game

Overview

This project contains three major files, AI.py file creates the model and train the model to play the game, play_ai.py uses the saved pre-trained model to play the game. If you like play the game by your own you can play using game.py script.

Installation Dependencies:

• Pygame
• Keras
• Tensorflow or Theano
• Numpy

How to run ?

 git clone https://github.com/mynkpl1998/Save-me-AI.git
 cd Save-me-AI
 python play_ai.py

How to train ?

 git clone https://github.com/mynkpl1998/Save-me-AI.git
 cd Save-me-AI
 python AI.py

Network Architecture

• Convert the screen images to grayscale
• Stack the four frames together
• 3 Convolutional Neural Network without maxpool layers
• 2 Fully Connected Dense layers

Hyperparameters

• Learning Rate (Default : 1e-06)
• Initial Epsilon (Default : 0.1)
• Final Epsilon (Default : 0.0001)