https://github.com/yanpanlau/Keras-FlappyBird

Using Keras and Deep Q-Network to Play FlappyBird
https://github.com/yanpanlau/Keras-FlappyBird

Last synced: 3 months ago
JSON representation

Using Keras and Deep Q-Network to Play FlappyBird

• Host: GitHub
• URL: https://github.com/yanpanlau/Keras-FlappyBird
• Owner: yanpanlau
• Created: 2016-07-12T11:15:35.000Z (over 8 years ago)
• Default Branch: master
• Last Pushed: 2019-03-05T16:01:38.000Z (over 5 years ago)
• Last Synced: 2024-07-19T02:14:50.779Z (4 months ago)
• Language: Python
• Size: 30.8 MB
• Stars: 434
• Watchers: 31
• Forks: 194
• Open Issues: 11
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Keras-FlappyBird

A single 200 lines of python code to demostrate DQN with Keras

Please read the following blog for details

https://yanpanlau.github.io/2016/07/10/FlappyBird-Keras.html

![](animation1.gif)

# Installation Dependencies:

* Python 2.7

* Keras 1.0 

* pygame

* scikit-image

# How to Run?

**CPU only**

```

git clone https://github.com/yanpanlau/Keras-FlappyBird.git

cd Keras-FlappyBird

python qlearn.py -m "Run"

```

**GPU version (Theano)**

```

git clone https://github.com/yanpanlau/Keras-FlappyBird.git

cd Keras-FlappyBird

THEANO_FLAGS=device=gpu,floatX=float32,lib.cnmem=0.2 python qlearn.py -m "Run"

```

If you want to train the network from beginning, delete the model.h5 and run qlearn.py -m "Train"