Flappy Bird Reinforcement Learning (Deep Q-Learning)

This project implements a reinforcement learning agent that learns to play Flappy Bird using Deep Q-Learning.

Features

• Flappy Bird game implemented in Pygame
• Deep Q-Network (DQN) for training an AI agent
• Replay buffer for experience replay
• Epsilon-greedy exploration strategy
• Automatic model saving and loading

Installation

• git clone https://github.com/dylan-savage/flappy-bird-ML.git
• cd flappy-bird-ML
• pip install -r requirements.txt

Playing

• To play manually run: python main.py
• To train the agent: python agent.py