This repository contains a reinforcement learning (RL) implementation of the Flappy Bird game. The RL agent is trained using Deep Q-Learning, enabling it to learn and master the game by avoiding obstacles and surviving as long as possible.
- Deep Q-Learning: Utilizes a neural network to approximate the Q-value function.
- Flappy Bird Environment: The classic Flappy Bird game recreated in a Python environment.
- Training and Testing Modes: Allows training the RL agent from scratch or testing a pre-trained model.
- Customizable Parameters: Easy to modify hyperparameters like learning rate, discount factor, and more.
To get started with this project, follow these steps:
-
Clone the repository:
git clone https://github.com/Soham-KT/Flappy-Bird-RL.git cd Flappy-Bird-RL -
Install Dependencies:
pip install -r requirements.txt
You can customize various aspects of the training process by modifying the config.py file. This includes:
- Learning rate
- Discount factor (gamma)
- Exploration strategy (epsilon decay)
- Batch size
- And more...
Contributions are welcome! If you'd like to contribute, please fork the repository and use a feature branch. Pull requests are warmly welcomed.