Python based agent that learns to play Flappy bird
A single 200 lines of python code to demostrate DQN with Keras
-# Installation Dependencies: -* Python 2.7 -* Keras 1.0 -* Tensorflow 1.0+ -* pygame -* scikit-image -* scipy and numpy
If you want to run the original pre-trained network,
- -python qlearn_orig.py -
If you want to train the network from beginning, delete the model.h5 and run
- -python qlearn.py -m "Train -
Reference: Please read the following blog for details https://yanpanlau.github.io/2016/07/10/FlappyBird-Keras.html