This is a demo of buidling a simple neural network with 2 layers a hidden layer and an output layer. The hidden layer will use the sigmoid function for activations.
usign only numpy only and then training and testing mnist benchmark data
we were able to achieve 97% accuracy with this simple neural network on mnist data, This shows power of simple neural network over sophisticated machine learning algorithms.
huge thanks to Joseph Redmon for providing mnist data in csv format http://pjreddie.com/projects/mnist-in-csv/