Open AI games
This is a place for me to play around with different reinforcement learning algorithms and explore their capability to solve various scenarios as provided from the OpenAI gym environment.
In addition to SciPy, Numpy, the tensorflow python library, openCV and matplotlib there are several other requirements that cannot be installed through Pip or Conda.
Installing tensorflow: https://www.tensorflow.org/install/install_windows
Getting CUDA v9 (required for tensorflow gpu use): https://developer.nvidia.com/cuda-90-download-archive
Getting cuDNN: https://docs.nvidia.com/deeplearning/sdk/cudnn-install/index.html
Download and store in datasets/housenumbers. There are 3 files, grab each of them. They are MATLAB style .mat files. http://ufldl.stanford.edu/housenumbers/