https://github.com/harishtpj/rawnn

A minimal Neural Network framework written from scratch
https://github.com/harishtpj/rawnn

framework neural-network numpy python

Last synced: 6 days ago
JSON representation

A minimal Neural Network framework written from scratch

• Host: GitHub
• URL: https://github.com/harishtpj/rawnn
• Owner: harishtpj
• License: mit
• Created: 2026-05-15T08:37:18.000Z (about 1 month ago)
• Default Branch: master
• Last Pushed: 2026-05-15T13:49:25.000Z (about 1 month ago)
• Last Synced: 2026-05-15T15:35:18.717Z (about 1 month ago)
• Topics: framework, neural-network, numpy, python
• Language: Python
• Homepage:
• Size: 6.84 KB
• Stars: 0
• Watchers: 0
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# 🤖 RawNN

A minimalistic, lightweight Neural Network module for quick prototyping in Python.

# ℹ️ About

**RawNN** is a modular, toy neural network builder written entirely from scratch in Python using only 

NumPy. No TensorFlow, no PyTorch; just raw math and logic.

The project explores the fundamental building blocks of Deep Learning, implementing everything from 

individual Neurons to Convolutional Neural Networks (CNNs). It provides a similar Tensorflow/Keras-like 

API for familiarity and uses a Neuron-centric model.

Note that this project doesn't aim on efficiency, or faster performance; rather it focuses on giving a

simple implementation for users to understand how the network works on backend.

# 📃 Usage

Clone this repository and ensure that you've NumPy installed

```bash

git clone [https://github.com/harishtpj/RawNN.git](https://github.com/harishtpj/RawNN.git)

pip install numpy

```

Example models implemented for `iris` and `mnist` datasets are provided in the root of repository.

You may need to install `scikit-learn` for running those (for downloading the respective datasets).

# 📝 License

#### Copyright © 2026 [M.V.Harish Kumar](https://github.com/harishtpj). 


#### This project is [MIT](https://github.com/harishtpj/RawNN/blob/master/LICENSE) licensed.