# mattjj/autodidact

A pedagogical implementation of Autograd

Repository: https://github.com/mattjj/autodidact
Canonical: https://ross.abutalabs.com/products/autodidact
Language: Jupyter Notebook
License: MIT
License Family: permissive
Last push: 2020-05-26T14:11:59+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3137, "days_push": 2290, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1024, forks 108 (observed 2026-08-28T04:03:16.403510+00:00)

## What it is
A small, tutorial-style reimplementation of the Autograd automatic differentiation library, written for learning how reverse-mode autodiff works. It provides the same grad-based API as Autograd over a thinly-wrapped NumPy.

## Use cases
- learn how automatic differentiation works internally
- understand reverse-mode gradient computation
- compute gradients and higher-order derivatives of numpy functions
- study the source of a minimal autograd system
- teach a course on differentiable programming

## When to choose
- you want readable source code explaining how autodiff is implemented
- you are teaching or learning the internals of tools like Autograd or JAX
- you need a tiny dependency-free autodiff to experiment with

## When to avoid
- you need a production-grade or performant autodiff system
- you want GPU acceleration or large-scale deep learning support
- you need an actively maintained library - use JAX or PyTorch instead

## Facets
- artifact type: learning-resource
- maturity: maintenance
- function: machine-learning, math, developer-tools
- domain: machine-learning, education
- platform: python
- tags: autodiff, automatic-differentiation, autograd, pedagogical, jupyter-notebook, gradients, algorithms

## Member repositories
- mattjj/autodidact (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:16.403510+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 2026-08-30T07:08:14.906160+00:00, confidence not recorded.
  - readme: https://github.com/mattjj/autodidact (fetched 2026-08-28T04:03:16.403510+00:00, sha e0d96b60a905)
- Data as of 2026-08-30T08:39:29.467469+00:00.