Siwoo4985/Micro-Diffusion

The essence of text diffusion in ~150 lines of pure Python. Inspired by Karpathy's MicroGPT.

32
/ 100
Emerging

This project provides a minimal, pure Python implementation to understand how text diffusion models work for generating text. It takes a list of short text entries, like names, and demonstrates how a model learns to 'denoise' them from a fully masked state. It's designed for machine learning practitioners and researchers who want to grasp the core mechanics of discrete text diffusion without complex frameworks.

Use this if you are a machine learning practitioner, student, or researcher looking for a simple, clear, and runnable example to understand the internals of text diffusion for discrete data.

Not ideal if you need a production-ready system for high-quality text generation, a library for integrating into a larger application, or a tool for large-scale training with massive datasets.

text-generation-algorithms natural-language-processing-education machine-learning-research discrete-diffusion-models
No Package No Dependents
Maintenance 10 / 25
Adoption 7 / 25
Maturity 11 / 25
Community 4 / 25

How are scores calculated?

Stars

27

Forks

1

Language

Python

License

MIT

Last pushed

Mar 12, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/Siwoo4985/Micro-Diffusion"

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