Microsoft is taking over PyTorch for Windows from Facebook – WindowsReport.com


Loading Comments

Download PDF


  • Facebook announced that Microsoft has taken ownership of the development and maintenance of the PyTorch build for Windows.
  • For those unaware, PyTorch is a machine learning library for Python programming language, more accessible than other deep learning tools.
  • If you want to learn more about developer tools head over to our Developer Tools section for a lot of interesting articles.
  • Are you interested in a more broad approach? Visit our Software Hub and learn about the latest tools.

Facebook announced that Microsoft has taken ownership of the development and maintenance of the PyTorch build for Windows.

For those unaware, PyTorch is a machine learning library for Python programming language, more accessible than other deep learning tools.

The move is an effort by Microsoft to improve the library’s performance on Windows 10 computers and the Windows Subsystem for Linux (WSL) a preferred tool for developers to run Linux distributions.

PyTorch, was released by Facebook in January 2017, and it was based on Torch, a scientific computing framework, and script language.

The tool helps developers use machine-learning Python packages and supports GPU-accelerated computations for data tasks.

PyTorch is seen by experts as one of the fastest-growing projects on GitHub, hence the close interest from Microsoft.

As for Facebook, they used PyTorch in the artificial intelligence tool Translate that powers translation for 48 languages on Facebook. 

According to the announcement, Windows support for PyTorch has lagged behind other platforms.

Lack of test coverage resulted in unexpected issues popping up every now and then. Some of the core tutorials, meant for new users to learn and adopt PyTorch, would fail to run.

The installation experience was also not as smooth, with the lack of official PyPI support for PyTorch on Windows.

Lastly, some of the PyTorch functionality was simply not available on the Windows platform, such as the TorchAudio domain library and distributed training support.

PyTorch will improve on WSL and Nvidia’s CUDA platform

As we mentioned above, moving PyTorch to Microsoft will also help to improve WSL on Windows 10, which has preview support for GPU-accelerated machine-learning (ML) training.

And that opens the door for developers and data scientists to use Nvidia’s CUDA platform to accelerate training.

Are you a developer? How do you feel about this move? Tell us all about it in the Comments section below.