---
id: autograd-gamma
version: "0.5.0"
license: MIT License
license_treatment: permissive
maintenance: abandoned
---

# autograd-gamma — Autograd compatible approximations to the gamma family of functions

License: permissive · Maintenance: abandoned · Downloads: 1.9M/mo

## What it is and what it does

autograd-gamma extends the autograd automatic differentiation library with derivatives for incomplete gamma and beta functions—special mathematical functions commonly used in statistics and scientific computing. It wraps approximations to these derivatives so you can compute gradients through them in optimization and machine learning workflows.

The package provides four gamma-family functions (gammainc, gammaincc, gammaincln, gammainccln) and two beta-family functions (betainc, betaincln), each differentiable with respect to specified arguments. It has no runtime dependencies and is designed as a lightweight bridge. However, the project has been inactive since 2020-10-15 and is no longer maintained.

Use it for:

- Computing gradients through incomplete gamma functions in Bayesian inference or statistical models
- Optimizing parameters in distributions that depend on gamma or beta functions using gradient descent
- Building automatic differentiation pipelines for scientific computing that require special function derivatives

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Provides autograd-compatible derivatives for incomplete gamma and beta functions, enabling automatic differentiation through special mathematical functions.

No, unless you have a specific legacy dependency. The package is abandoned with last commit 2020-10-15 and no maintenance signal. High install friction and uncertain compatibility with current autograd and Python versions make it a poor choice for new projects.

## Install

pip install autograd-gamma
uv add autograd-gamma
poetry add autograd-gamma

## Installing autograd-gamma

Before you install: High install friction and abandoned status since 2020-10-15 with no updates since. No runtime dependencies, but lack of maintenance may cause compatibility issues with modern Python versions.

License in practice: MIT License permits free use, modification, and distribution with minimal restrictions.

Quickstart:

from autograd import grad
from autograd_gamma import gammainc, gammaincc

grad(gammainc, argnum=0)(1., 2.)
grad(gammaincc, argnum=0)(1., 2.)

Requires autograd to be installed separately; package is unmaintained since 2020-10-15

Verify before relying:

- Compatibility with current versions of autograd and modern Python releases
- Whether the analytical solutions mentioned in the long-term goal were completed or merged upstream
- Performance characteristics of the approximations relative to alternatives

## Package facts

- License: MIT License (permissive)
- Python support: unspecified
- Install friction: high
- Maintenance: abandoned
- Downloads: 1.9M/month (top 5,000 on PyPI)
- Known vulnerabilities: none known

## Tags

autograd gamma function derivatives, incomplete gamma function autodiff, special functions automatic differentiation, beta function gradients, scientific computing derivatives, autodiff, special-functions, abandoned

[View on SkillFed](https://skillfed.io/packages/autograd-gamma) · [View on PyPI](https://pypi.org/project/autograd-gamma/)