anh-nn01/Neural-Network-from-Scratch--Hand-written-Digits-classifier

This is my first Deep Learning project, which is a MNIST hand-written digits classifier. The model is implemented completely from scratch WITHOUT using any prebuilt optimization like Tensorflow or Pytorch. Tensorflow is imported only to load the MNIST data set. This model also uses 2 hidden layers with Adaptive Moment Optimization (Adam) and Drop-out regularization.

13
/ 100
Experimental

This project helps anyone interested in the foundational mechanics of deep learning to understand how neural networks classify images. It takes raw image data, specifically handwritten digits, and processes it through a custom-built neural network to output predicted digit labels. It's designed for someone looking to grasp the underlying mathematics and computational steps of deep learning.

No commits in the last 6 months.

Use this if you are a student or educator wanting to learn or teach the fundamental, 'from-scratch' implementation of a neural network for image classification.

Not ideal if you need a high-performance, production-ready image classifier or prefer using established deep learning frameworks like TensorFlow or PyTorch directly.

deep-learning-education neural-networks image-classification machine-learning-fundamentals algorithm-implementation
No License Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 5 / 25
Maturity 8 / 25
Community 0 / 25

How are scores calculated?

Stars

9

Forks

Language

Jupyter Notebook

License

Last pushed

Jun 13, 2020

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/anh-nn01/Neural-Network-from-Scratch--Hand-written-Digits-classifier"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.