A personalised exercise on Neural Networks following the book http://neuralnetworksanddeeplearning.com/ by michael nielsen
Instead of the SDG (Stochastic Gradient Descent) I used the usual gradient descent algo. I wrote the code 1 year ago more or less, before learning that is a very useful practice to comment your work. Anyway the records are quite easy to read anyway, and report data on the experiment without the need for a legend, most of times.