# tanelp/tiny-diffusion

A minimal PyTorch implementation of probabilistic diffusion models for 2D datasets.

Repository: https://github.com/tanelp/tiny-diffusion
Canonical: https://ross.abutalabs.com/products/tiny-diffusion
Language: Jupyter Notebook
License Family: other
Last push: 2024-05-07T18:32:13+00:00

## Health v2 (maintenance only)
Score: 31/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 94
- inputs: {"age_days": 1328, "days_push": 848, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1014, forks 81 (observed 2026-08-28T04:03:13.820889+00:00)

## What it is
A minimal PyTorch implementation of probabilistic diffusion models (DDPM) trained on 2D point datasets like the Datasaurus Dozen. It includes visualizations of the forward and reverse diffusion processes plus ablation experiments on hyperparameters such as learning rate, timesteps, and variance schedules.

## Use cases
- learn how diffusion models work
- understand ddpm training internals
- experiment with diffusion model hyperparameters
- visualize forward and reverse diffusion processes
- minimal diffusion model code to read and modify
- study effect of timestep count and beta schedules

## When to choose
- you want a small, readable codebase to learn diffusion models from
- you need a sandbox for running ablations on DDPM hyperparameters
- you prefer minimal implementations over large production frameworks

## When to avoid
- you need production-ready image or video diffusion generation
- you want pretrained models or a full feature set like HuggingFace diffusers
- you need a maintained library with a license and long-term support

## Facets
- artifact type: learning-resource
- maturity: maintenance
- function: machine-learning, deep-learning, data-visualization
- domain: deep-learning, machine-learning, tutorials, data-visualization
- platform: python
- tags: diffusion-models, ddpm, pytorch, educational, generative-models, minimal-implementation

## Member repositories
- tanelp/tiny-diffusion (main) score 31

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:13.820889+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:11:35.886872+00:00, confidence not recorded.
  - readme: https://github.com/tanelp/tiny-diffusion (fetched 2026-08-28T04:03:13.820889+00:00, sha 68422a54372c)
- Data as of 2026-08-30T08:39:29.467469+00:00.