workofart/ml-by-hand
A deep learning library built from scratch with complex neural networks examples built on top for learning purposes.
This project is for software developers and machine learning engineers who want to deeply understand how neural networks work from the ground up. It lets you build and train deep learning models like GPT-2 by hand, exposing all the underlying mathematical details. You input raw data and model configurations, and it outputs a working neural network, helping you grasp concepts often hidden by high-level libraries.
Use this if you are a developer looking to learn the fundamental mathematics and algorithms behind deep learning by implementing them yourself.
Not ideal if you need a high-performance, production-ready deep learning framework or prefer to work with existing abstractions to quickly build models.
Stars
76
Forks
14
Language
Python
License
MIT
Category
Last pushed
Mar 10, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/workofart/ml-by-hand"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related frameworks
uxlfoundation/scikit-learn-intelex
Extension for Scikit-learn is a seamless way to speed up your Scikit-learn application
INRIA/scikit-learn-mooc
Machine learning in Python with scikit-learn MOOC
ddbourgin/numpy-ml
Machine learning, in numpy
nubank/fklearn
fklearn: Functional Machine Learning
gavinkhung/machine-learning-visualized
ML algorithms implemented and derived from first-principles in Jupyter Notebooks and NumPy