Neural-Network-Python

Feed-forward Neural network implementation in python - using numpy library.

Functions Usage

• Activation Values

  The file contain a function which will return the activation values in output of an input layer

• Back Propagation

  The file contain a function which will return a tuple as the gradient for the cost function, according to the famous back propagation algorithm.

• Cost Derivative

  This file contain a function which returns the vector of partial derivatives.

• Sigmoid Functions

  This file contains 2 functions:

sigmoid.py

It returns the sigmoid function

sigmoid_prime

It returns the derivative of sigmoid function

• Stochastic Gradient Descent

  This File contains a function which will train Neural Network using mini-batchs.

• Update Mini Batch

  The File contains a function which will update the network's weights and biases by applying gradient descent using backpropagation to a single mini batch.

Neural Network

This File contains a Class 'Network' which consists of all the above mentioned functions and appropriate function calls.

Thanks to -

• Numpy library
• Michael Nielson