Investigating Neural Networks
Hunter Herman, Joweina Hsiao, Talha Ahsan

Welcome to our project!
We provide an installation script "setup.sh" that will make sure the environment is prepared for running our code. After you run the script, make sure you also run `source bin/activate` so that the installed packages are available. Alternately, if you'd rather set it up by yourself, make sure you've done the following:
-installed pybrain (pip install pybrain)
-downloaded the MNIST dataset

Regardless of how you complete the prior steps, you MUST have done the following:
-adjusted the various PATH variables in writeNetworks.py (at the bottom) to suit your system

If you'd like to try the tensorflow implementation of Convolutional Neural Networks, install tensorflow as well! env.sh will do this for you, but otherwise you can simply run:
	pip install --upgrade https://storage.googleapis.com/tensorflow/mac/tensorflow-0.5.0-py2-none-any.whl
WARNING: This is the Mac/CPU-only version of Tensorflow, if you'd like to try a different installation, visit their website tensorflow.org.

Finally, you should be ready to run our project. Any of the following will produce some interesting results:
python writeNetworks.py
python tensor.py

writeNetworks.py will train and evaluate multiple Convolutional Neural Networks, LSTM networks, and Feed Forward Networks. tensor.py will train and evaluate a single 2-layer convolutional Neural Network using the Tensorflow backend.