标签

 modular 

相关的文章:

Modular: Developer Voices, Deep Dive, Extensibility, Supply Chain Attacks, Performance Analysis and more. Explore the latest articles on modular development and discover insights from industry experts.

Modular: What Does Joe Pamer, AI and PL expert, Want From Mojo?

原文英文,约1700词,阅读约需7分钟。发表于:

What Does Joe Pamer, AI and PL expert, Want From Mojo?

Joe Pamer, a veteran of Meta, Apple, and Microsoft, has joined Modular to lead the Mojo engineering team. He believes Mojo is a balanced solution for machine learning pipelines, combining the power of C++ and the expressiveness of Python. Pamer aims to improve tooling, grow the Mojo ecosystem, and sees potential for Mojo as a general-purpose programming language. He is excited about the tech scene in New York City and grateful for the opportunity to work remotely for Modular. Pamer also advocates for the use of emojis in programming languages.

Modular: What Does Joe Pamer, AI and PL expert, Want From Mojo?
相关推荐 去reddit讨论

Modular: What’s New in Mojo 24.3: Community Contributions, Pythonic Collections and Core Language Enhancements

原文英文,约2200词,阅读约需8分钟。发表于:

What’s New in Mojo 24.3: Community Contributions, Pythonic Collections and Core Language Enhancements

Mojo🔥 24.3是自开源Mojo🔥标准库以来的首个重大版本,包含了许多新功能和改进,特别是对集合类型的增强,以及梯度下降算法和集合操作的代码示例。此外,还介绍了Set集合的新方法和reversed()函数的使用。整个版本还包含了其他一些核心语言和标准库的改进。

Modular: What’s New in Mojo 24.3: Community Contributions, Pythonic Collections and Core Language Enhancements
相关推荐 去reddit讨论

Modular: MAX 24.3 - Introducing MAX Engine Extensibility

原文英文,约1200词,阅读约需5分钟。发表于:

MAX 24.3 - Introducing MAX Engine Extensibility

MAX 24.3发布,包括MAX Engine Extensibility API预览,允许开发者在新一代编译器和运行时堆栈上统一、编程和组合AI流水线,以实现最佳性能。MAX Engine是一个用于运行AI推理的新一代编译器和运行时库,支持PyTorch(TorchScript)、ONNX和本地Mojo模型,可在各种硬件上实现低延迟、高吞吐量的推理。此外,MAX平台通过使用MAX Graph API创建定制推理模型,使您能够充分发挥MAX Engine的潜力。MAX 24.3的新功能包括自定义操作,使构建AI流水线时更容易使用自定义操作。通过MAX Graphs和Mojo,您可以编写自己的MAX Graphs,并使用自定义操作扩展。MAX Engine和Mojo建立在MLIR之上,充分利用现代编译器的优势。MAX Engine确保您的Mojo代码在各种平台和硬件设置上可移植,扩大了部署选项。

Modular: MAX 24.3 - Introducing MAX Engine Extensibility
相关推荐 去reddit讨论

Modular: Developer Voices: Deep Dive with Chris Lattner on Mojo

原文英文,约6400词,阅读约需24分钟。发表于:

Developer Voices: Deep Dive with Chris Lattner on Mojo

Mojo是一个基于MLIR编译器框架的代码生成器,现在是Python的超集,提供高性能和可移植性。它的类型系统具有静态和动态行为,内存管理和可变性特性受到Rust和Swift的影响。该语言支持并行化、CPU和GPU的组合,并旨在弥合人工智能研究和生产之间的差距。Mojo的开源社区正在迅速发展,整个标准库正在开源化。

Modular: Developer Voices: Deep Dive with Chris Lattner on Mojo
相关推荐 去reddit讨论

Modular: Preventing Supply Chain Attacks at Modular

原文英文,约1100词,阅读约需4分钟。发表于:

Preventing Supply Chain Attacks at Modular

Modular focuses on secure software delivery and defends against vulnerabilities in the supply chain. They use The Update Framework (TUF) for security and offer the Modular CLI for secure software distribution. They plan to expand their client and hire new employees.

Modular: Preventing Supply Chain Attacks at Modular
相关推荐 去reddit讨论

Modular: Row-major vs. Column-major Matrices: A Performance Analysis in Mojo and NumPy

原文英文,约3000词,阅读约需11分钟。发表于:

Row-major vs. Column-major Matrices: A Performance Analysis in Mojo and NumPy

本文介绍了矩阵在计算机内存中的存储方式,行主序和列主序的区别,以及它们在不同算法和编程语言中的性能差异。文章还介绍了使用MojoMatrix在Mojo中实现行主序和列主序矩阵的方法,并比较了它们与NumPy的性能差异。基准测试展示了行主序和列主序矩阵在性能上的差异,列主序矩阵在某些情况下可以比行主序矩阵快几倍。

Modular: Row-major vs. Column-major Matrices: A Performance Analysis in Mojo and NumPy
相关推荐 去reddit讨论

Modular: How to Be Confident in Your Performance Benchmarking

原文英文,约2700词,阅读约需10分钟。发表于:

How to Be Confident in Your Performance Benchmarking

Mojo是一种性能强大、可编程性高、可移植的语言。它能够让用户编写快速的代码,并且易于使用,可在不同的CPU平台上运行。通过遵循基准测试原则,可以确保性能结果的准确性。

Modular: How to Be Confident in Your Performance Benchmarking
相关推荐 去reddit讨论

Modular: The Next Big Step in Mojo🔥 Open Source

原文英文,约1000词,阅读约需4分钟。发表于:

The Next Big Step in Mojo🔥 Open Source

Modular has open sourced the core modules of the Mojo standard library under the Apache 2 license, aiming for better collaboration with developers worldwide. They provide revision history, nightly builds, and public CI, and accept external contributions through GitHub pull requests. The Apache 2 license with LLVM exceptions is used to address concerns about mixing with GPL2 code. Modular plans to release more code in the future and offers swag and benefits to contributors. Nightly builds of the Mojo compiler and version control history of the standard library are available. They are excited about the future of Mojo and welcome contributions and feedback.

Modular: The Next Big Step in Mojo🔥 Open Source
相关推荐 去reddit讨论

Modular: How to Contribute to Mojo Standard Library: A Step-by-Step Guide

原文英文,约1900词,阅读约需7分钟。发表于:

How to Contribute to Mojo Standard Library: A Step-by-Step Guide

最近,我们宣布了Mojo标准库的开源。这是我们社区的一个重要里程碑,不仅提供了了解标准库各种功能实现细节的最佳方式,还为贡献到Mojo创造了绝佳机会。自宣布以来,我们幸运地收到了各种贡献,从文档修复到代码更改。为了更好地促进社区的贡献,本文将逐步指导您如何贡献到Mojo标准库,包括找到GitHub问题、Git设置、编写代码、进行本地测试以及创建拉取请求(PR)并合并。请参考我们的贡献指南,其中包含详细信息,包括Mojo行为准则和样式指南,本文未涵盖。贡献到Mojo可以采取多种形式,从参与讨论和识别或报告问题,到通过RFC提出语言更改。在这个背景下,我们将重点关注代码贡献,这涉及一系列我们将在下面介绍的步骤。

Modular: How to Contribute to Mojo Standard Library: A Step-by-Step Guide
相关推荐 去reddit讨论

Modular: What’s new in Mojo 24.2: Mojo Nightly, Enhanced Python Interop, OSS stdlib and more

原文英文,约2100词,阅读约需8分钟。发表于:

What’s new in Mojo 24.2: Mojo Nightly, Enhanced Python Interop, OSS stdlib and more

上周我们发布了MAX 24.2版本,详细介绍了MAX 24.2的新功能。本文将深入介绍Mojo SDK 24.2的所有新功能,特别是针对Python开发者。新功能包括Mojo标准库的开源、DynamicVector更名为List、List支持不同类型的元素、math模块新增ulp()函数、Mojo与Python的互操作性、新的核心语言特性等。此外,还提供了代码示例和Jupyter Notebook供参考。

Modular: What’s new in Mojo 24.2: Mojo Nightly, Enhanced Python Interop, OSS stdlib and more
相关推荐 去reddit讨论