MNIST Classification with a Deep CNN
Simply run code by calling python CSC416Program.py from the command line.
There are a number of arbitrarily chosen values within the program that you can feel free to change and toy with. They're all just stored globally at the top of the file. FILE_NAME sets the tmp/FILE_NAME directory where TensorFlow stores checkpoints for the network. To start training again from scratch, either clear out that file, or change FILE_NAME to an unused directory path.
A NIPS formatted paper can be found at Paper/nips_2018.tex within this repository.
TensorFlow can pull up the MNIST dataset without the user downloading it themselves. That being said, if you would like access to the dataset for any other reason, you can find it here: http://yann.lecun.com/exdb/mnist/