arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.08770astro-ph.IMastro-ph.GAphysics.comp-ph
收藏

unxt:基于JAX的单位感知计算Python包

unxt: A Python package for unit-aware computing with JAX

Center for Computational Astrophysics, Flatiron Institute(计算天体物理中心,平顿研究所);Department of Physics, Princeton University(物理学系,普林斯顿大学);Kavli Institute for Astrophysics and Space Research, Massachusetts Institute of Technology(卡弗里天体物理学与空间研究所,麻省理工学院)

Nathaniel Starkman, Adrian Price-Whelan, Jake Nibauer

AI总结:

unxt通过集成astropy.units实现单位感知计算,提升JAX在科学计算中的性能与准确性。

摘要:

unxt是一个基于JAX的Python包,用于单位感知计算。unxt建立在quax之上,后者提供了一个构建数组样对象的框架,这些对象可以与JAX一起使用。unxt扩展了quax,以提供使用astropy.units包作为单位后端的单位感知计算支持。unxt提供了将物理单位无缝集成到高性能数值计算中的能力,显著增强了JAX在科学应用中的功能。

英文摘要:

unxt is a Python package for unit-aware computing with JAX. unxt is built on top of quax, which provides a framework for building array-like objects that can be used with JAX. unxt extends quax to provide support for unit-aware computing using the astropy.units package as a units backend. unxt provides seamless integration of physical units into high performance numerical computations, significantly enhancing the capabilities of JAX for scientific applications.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑