Metadata-Version: 2.1
Name: gustavgrad
Version: 0.1.1a0
Summary: 
Home-page: https://github.com/gustavgransbo/gustavgrad
License: MIT
Author: Gustav Gränsbo
Author-email: gustav.gransbo@gmail.com
Requires-Python: >=3.7,<4.0
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.7
Classifier: Programming Language :: Python :: 3.8
Requires-Dist: numpy (>=1.19.0,<2.0.0)
Project-URL: Repository, https://github.com/gustavgransbo/gustavgrad
Description-Content-Type: text/markdown

# gustavgrad
[![Tests](https://github.com/gustavgransbo/gustavgrad/workflows/Tests/badge.svg)](https://github.com/gustavgransbo/gustavgrad/actions?workflow=Tests)
[![codecov](https://codecov.io/gh/gustavgransbo/gustavgrad/branch/master/graph/badge.svg)](https://codecov.io/gh/gustavgransbo/gustavgrad)

An autograd library built on NumPy, inspired by [Joel Grus's livecoding](https://github.com/joelgrus/autograd/tree/master).

The idea behind gustavgrad is to define a Tensor class, and a set of arithmetic operations on tensors, which we know how to calculate the first order derivative for.
Using the chain-rule, the gradient of the composition of multiple operations can be calculated, since we know how to calculate the first order derivative of the basic operations.

