Neural Networks from Scratch (NumPy)

This project implements neural networks from scratch using only Python and NumPy, without deep learning frameworks like TensorFlow or PyTorch.
The models are trained to classify images as cat or non-cat— includes a 2-layer and an L-layer model trained to classify cat vs non-cat images.

📌 Features

• 2-layer neural network implementation
• L-layer (deep) neural network implementation
• Forward and backward propagation from scratch
• Uses sigmoid, ReLU activations
• Gradient descent optimization
• Trained on "cat vs non-cat" dataset

Dataset

The dataset files are included in this repository:

• train_catvnoncat.h5
• test_catvnoncat.h5

Requirements

• Python 3.x
• NumPy