https://github.com/sidsurakanti/mnist-classification

mnist digit classification from scratch w/ only numpy and maths
https://github.com/sidsurakanti/mnist-classification

math ml mnist numpy

Last synced: 10 months ago
JSON representation

mnist digit classification from scratch w/ only numpy and maths

• Host: GitHub
• URL: https://github.com/sidsurakanti/mnist-classification
• Owner: sidsurakanti
• Created: 2025-04-12T03:29:25.000Z (10 months ago)
• Default Branch: main
• Last Pushed: 2025-04-14T23:29:26.000Z (10 months ago)
• Last Synced: 2025-04-15T00:29:31.286Z (10 months ago)
• Topics: math, ml, mnist, numpy
• Language: Jupyter Notebook
• Homepage:
• Size: 2.12 MB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
## Overview

MNIST digit classifier built with just NumPy and Math. It also has a Pytorch + a realtime web implementation using FastAPI and Next.js. 



  

    

    

  

## What I learned

- How to build a neural net from scratch  

- Forward and backward propagation using pure math

- Vectorized operations with NumPy  

- Loss functions, softmax, and gradient descent

- Data normalization and preprocessing 

- Pytorch

- Model deployment with FastAPI

## Stack

![Python](https://img.shields.io/badge/python-%2314354C.svg?style=for-the-badge&logo=python&logoColor=white)

![NumPy](https://img.shields.io/badge/numpy-%23013243.svg?style=for-the-badge&logo=numpy&logoColor=white)

![Matplotlib](https://img.shields.io/badge/matplotlib-2067b8?style=for-the-badge&logo=matplotlib&logoColor=white)

![Pandas](https://img.shields.io/badge/pandas-150458?style=for-the-badge&logo=pandas&logoColor=white)

![PyTorch](https://img.shields.io/badge/pytorch-%23ee4c2c.svg?style=for-the-badge&logo=pytorch&logoColor=white)

![FastAPI](https://img.shields.io/badge/fastapi-005571?style=for-the-badge&logo=fastapi)

![Next.js](https://img.shields.io/badge/next.js-000000?style=for-the-badge&logo=next.js&logoColor=white)

![TailwindCSS](https://img.shields.io/badge/tailwindcss-%2338bdf8.svg?style=for-the-badge&logo=tailwind-css&logoColor=white)

## Getting started

Play around with the project w/ demo.py or view the demo web app locally.

### Prerequisites

- Python 3.x  

- pip

- Node.js 18+

- npm

### Installation

```bash

git clone https://github.com/sidsurakanti/mnist-digit-recog.git

cd /path/to/project/

# backend setup

pip install -r requirements.txt

fastapi dev predict.py

# frontend setup

cd frontend

npm install

npm run dev

```

The app should now be live on http://localhost:3000!

## Roadmap

- [x] Build a neural net with only NumPy  

- [x] Implement backpropagation 

- [X] Add softmax + cross-entropy   

- [x] Visualize accuracy/loss  

- [X] Rebuild using Pytorch

- [X] Add GUI to draw digits and classify in real-time 

- [ ] Deploy online

- [ ] Improve digit preprocessing for the HTML canvas