This is a Pytorch Implementation of DQN for Playing Flappy Bird, including Q-learning and DQN with different types of neural networks for function approximation
- Q-Learning (Tabular)
- DNN
- CNN
- RNN
The usage of flappy bird game environment is adopted from https://github.com/taivu1998/FlapAI-Bird.
Flappy Bird game environment
git clone https://github.com/ntasfi/PyGame-Learning-Environment.git
cd PyGame-Learning-Environment/
pip install -e .
Other packages
pip install gym==0.15.4
pip install pygame
pip install gym_ple
pip install opencv-python
For Pytorch, one can refer to the official Pytorch documentation for installation