Neural Network Implementation in Python

This project showcases a simple implementation of a MNIST NN in Python using only vanilla Python and NumPy, focusing on understanding the fundamental concepts and operations involved. The neural network is designed to learn from the given training data and make predictions accordingly.

Usage

To use this implementation, create a Network instance with the desired architecture and training parameters. Then, utilize the SGD method to train the network using your training data.

# Create a neural network with desired architecture
net = Network([input_size, hidden_layer_size, output_size])

# Train the network using Stochastic Gradient Descent
net.SGD(training_data, epochs, mini_batch_size, eta, test_data=test_data)

Acknowledgments

This project was inspired by and based on the works of:

• GeeksforGeeks - Implementation of Neural Network from Scratch using NumPy
• Real Python - Python AI: Neural Network