neural-network-numpy

Implementation of a neural network from scratch using NumPy for MNIST digit classification with backpropagation and gradient descent.

Neural Network from Scratch (NumPy)

🔗 Project Link

https://github.com/Bitu-Singh-Rathoud/neural-network-numpy

📌 Overview

This project implements a neural network from scratch using NumPy without relying on high-level deep learning libraries. It demonstrates core concepts like forward propagation, backpropagation, and gradient descent for digit classification.

🚀 Features

• Fully connected neural network implementation
• Forward propagation and backpropagation
• Gradient descent optimization
• Activation functions (ReLU, Sigmoid, Softmax)
• Training on MNIST dataset

🛠 Tech Stack

• Python
• NumPy
• Machine Learning
• Deep Learning Fundamentals

📊 Results

• Achieved classification accuracy on handwritten digit dataset (MNIST)
• Demonstrated learning through loss reduction over epochs
• Visualized prediction outputs and performance

▶️ How to Run

pip install -r requirements.txt
python main.py