This is my code to follow and experiment with Nielsen's image-based optical character recognition tutorial
These scrips are updated ones from the neuralnetworksanddeeplearning.com gh repo in order to work with Python 3.5.2
The testing file (test.py) contains all three networks (network.py, network2.py, network3.py) from the book and it is the starting point to run (i.e. train and evaluate) them.
In test.py there are examples of networks configurations with proper comments. I did that to relate with particular chapters from the book.