Skip to main content

Implements parts of the Python Array API

Project description

array-adhoc

This an ad-hoc implementation of parts of the Python Array API backed by backends that do not themselves fully implement them.

The purpose is to use those backends with normalized array api code. For example, using this, you can use pytorch and the python arrays api using the same interface.

It's ad-hoc because I'm only implementating parts of the spec when I use them, and I do the implementations barebones.

This could easily improve with some good pull requests.

Wrappers Included: - torch

Handoffs Included: - cupy - numpy

It is easy to add a new wrapper or handoff, you just import the api functions in a new named file. If time is short, just make a handoff first, and then if something is missing or wrong or not compatible with old releases, improve it into a wrapper.

Usage:

import array_adhoc.torch as xp

array = xp.asarray([2,4,9])
print(xp.sqrt(array))

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

array-adhoc-0.0.1.tar.gz (4.0 kB view hashes)

Uploaded Source

Built Distribution

array_adhoc-0.0.1-py3-none-any.whl (3.6 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page