Implementation-of-Neural-Net-in-Numpy

A from-scratch implementation of a flexible neural network using only NumPy. This project demonstrates the inner workings of neural networks without relying on high-level deep learning libraries.

Features

• Configurable multi-layer neural network
• Support for different activation functions (ReLU, Sigmoid)
• Implementation of backpropagation algorithm
• Choice of optimizers (SGD, Adam)
• Training script with data generation and visualization