Coded my own neural network to recognise hand drawn characters from the MNIST dataset
Used the following resources to learn the theory: http://neuralnetworksanddeeplearning.com/index.html - all the code is my own 3Blue1Brown's YT video series
I am pretty confident my code is correct however it is poorly optimised and takes an incredibly long time to train. I would need to spend time making sure that all computations should be done with respect to layers, allowing me to utilise numpy's functions to speed things up.
If anyone can spot any errors or lines that could be imrpoved, that would be much appreciated!