metabochumpy

This is an active fork of chumpy, Matt Loper's autodifferentiation tool for Python.

The fork's goals are modest:

Keep the library working in current versions of Python and other tools.
Make bug fixes.
Provide API stability and backward compatibility with the upstream version.
Respond to community contributions.

Installation

Install the fork:

pip install metabochumpy

And import it just like the upstream library:

import chumpy as ch

Overview

Chumpy is a Python-based framework designed to handle the auto-differentiation problem, which is to evaluate an expression and its derivatives with respect to its inputs, by use of the chain rule.

Chumpy is intended to make construction and local minimization of objectives easier.

Specifically, it provides:

Easy problem construction by using Numpy’s application interface
Easy access to derivatives via auto differentiation
Easy local optimization methods (12 of them: most of which use the derivatives)

Usage

Chumpy comes with its own demos, which can be seen by typing the following:

import chumpy
chumpy.demo() # prints out a list of possible demos

Acknowledgements

Chumpy was created by Matthew Loper.

License

This project is licensed udner the MIT License.