Simple catch game DQN agent is implemented by TensorFlow
Original code is written by Torch7 and Keras
Keras code (written by Eder Santanas) : here
Torch7 code (written by SeanNaren): here
TensorFlow code written by solaris33: here
To train a model, run the TrainCatchGame.py script.
python trainModel.py
Play and Visualization implemented in iPython.
To run, type into terminal:
iPython notebook
go to this directory and run playGame.ipynb script in kernel