MS Ai Insider

Revolutionize Machine Learning on Windows: PyTorch 7 Launches Arm Native Builds for Enhanced Performance

Posted by

ailona

–

April 23, 2025

PyTorch has launched Arm native builds for Windows, allowing developers to harness the full power of Arm64 architecture without local compilation. With the release of PyTorch 7, users can now easily develop and test machine learning models on Windows Arm64 devices, enhancing performance for tasks like image classification and generative AI.

Exciting News: PyTorch Arm Native Builds for Windows

Developers, get ready! PyTorch has just launched Arm native builds for Windows, making machine learning easier and more efficient. This new feature is part of the PyTorch 7 release and supports Python 3.1 Now, you can leverage the full power of Arm64 architecture on your Windows devices.

What’s New?

Previously, developers had to compile PyTorch locally to support Windows Arm64. This update eliminates that hassle. With the new Arm native builds, you can develop, train, and test machine learning models directly on Arm-powered Copilot+ PCs.

“This unlocks the potential to leverage the full performance of Arm64 architecture on Windows devices.”

Major Updates: Getting Started

To get started, you’ll need a few prerequisites. It’s recommended to install MSVC and Rust Visual Studio Build Tools. Here’s a quick checklist:

Install Visual Studio with C++ Desktop development.
Ensure you have the latest VS 2022 C++ ARM64/ARM64EC build tools.
Install Python 3.19 from Python.org.

Once you have everything set up, installing PyTorch is straightforward. Use the following command:

pip install --extra-index-url https://download.pytorch.org/whl torch

What’s Important to Know?

While the Arm native builds are a significant step forward, some packages may not support Arm natively on Windows. However, PyTorch can compile dependencies from source code using MSVC and Rust. This flexibility allows you to install essential packages like NumPy and Safetensors easily.

“Developers can now start using stable Arm native PyTorch builds for Windows to create applications that leverage AI.”

Example Use Case: Stable Diffusion

One exciting application of the new builds is in generative AI, particularly with Stable Diffusion. You can use Arm native PyTorch binaries to create stunning images based on textual prompts. The example provided showcases how to generate images using a simple interface.

In conclusion, the availability of Arm native builds for PyTorch on Windows is a game-changer for developers. It simplifies the development process and unlocks the full potential of Arm architecture. Don’t miss out—give the native binaries a try today!

PyTorch 7 introduces Arm native builds specifically for Windows Arm64 devices.

Developers can now easily install PyTorch without the need for local compilation.

The new builds support Python 3.12, enhancing compatibility for developers.

Applications include image classification, natural language processing, and generative AI.

Installation recommendations include using Visual Studio Build Tools and creating a virtual environment.

From the Windows Blog