The repository contains a dataset of handwrtitten images from MNIST and an Artificial Neural Network on python that trains from the dataset and executes the algorithm on a mini batch of 10,000 test cases repeated over 30 generations to map the accuracy. The algorithm learns using a stochastic gradient descent to speed up learning.

Execute the following command in the Python shell.

training_data, validation_data, test_data = \
mnist_loader.load_data_wrapper()

Now we'll set up a Neural Network with 30 hidden neurons.

net = network.Network([784,30,10])

Finally, we'll use stochastic gradient descent to learn from the MNIST training_data over 30 epochs, with a mini-batch size of 10, and a learning rate of η=3.0

net.SGD(training_data, 30, 10, 3.0, test_data=test_data)

Note - python libraries used for this are - numpy scikit-learn scipy Theano