PyTorch Lightning: Difference between revisions

Revision as of 23:50, 5 June 2021

Revision as of 15:19, 8 September 2021

Line 10:

| website = https://www.pytorchlightning.ai/

}}

PyTorch Lightning is an open-source [[Python (programming language)|Python]] library that provides a high-level interface for [[PyTorch]], a popular [[deep learning]] framework.<ref>{{Cite web|url=https://github.com/PyTorchLightning/pytorch-lightning|title=GitHub - PyTorch Lightning|last=|first=|date=2019-12-01|website=|url-status=live|archive-url=|archive-date=|access-date=}}</ref> It is a lightweight and high-performance framework that organizes PyTorch code to decouple the research from the engineering, making deep learning experiments easier to read and reproduce. It is designed to create scalable deep learning models that can easily run on distributed hardware while keeping the models hardware agnostic. PyTorch Lightning provides the ultimate flexibility for any type of research, and reduces the engineering boilerplate required to implement state-of-the-art AI features.{{promotion-inline|{{subst:June 2021}}|date=June 2021}}

In 2019, Lightning was adopted by the [[Conference on Neural Information Processing Systems|NeurIPS]] Reproducibility Challenge as a standard for submitting PyTorch code to the conference.<ref>{{Cite web|url=https://reproducibility-challenge.github.io/neurips2019/resources/|title=Reproducibility Challenge @NeurIPS 2019|last=|first=|date=2019-12-01|website=NeurIPS|url-status=live|archive-url=|archive-date=|access-date=2019-12-01}}</ref>