🧠 numpy-neuralnet

A simple neural network built from scratch using NumPy.

This project demonstrates the core building blocks of a three-layer neural network trained on 2D synthetic datasets like flower pattern, circles, moons, etc. It also includes an interactive Streamlit app for visualization.

🚀 Features

• Forward & backward propagation using vectors
• Parameter updates via gradient descent
• Support for multiple synthetic datasets
• Decision boundary visualization
• Streamlit interface

Install dependencies:

pip install -r requirements.txt

Run the App

streamlit run app.py