Metadata-Version: 2.1
Name: smallpebble
Version: 0.1.1
Summary: Minimal automatic differentiation implementation in Python, NumPy.
Home-page: https://github.com/sradc/smallpebble
Author: Sidney Radcliffe
Author-email: sidneyradcliffe@gmail.com
License: Apache License 2.0
Platform: UNKNOWN
Classifier: Programming Language :: Python :: 3
Classifier: Operating System :: OS Independent
Requires-Python: >=3.7
Description-Content-Type: text/markdown
Requires-Dist: numpy (>=1.20.0)

# SmallPebble

[![](https://github.com/sradc/smallpebble/workflows/Python%20package/badge.svg)](https://github.com/sradc/smallpebble/commits/)

<br>
<p align="center">
<img src="pebbles.jpg" />
</p>
<br>

Minimal automatic differentiation implementation in Python, NumPy.


For an introduction to autodiff and the basic concepts of this framework, see:
https://sidsite.com/posts/autodiff/


Consider this a resource on autodiff, rather than a library you should use.
(Popular libraries are: JAX, PyTorch, TensorFlow...)


Features:
- Various operations, such as matmul, conv2d, maxpool2d.
- Supports broadcasting.
- Nth derivatives.


