By playing a more profound role in maintaining PyTorch, Microsoft will be able to make the library better support systems running on Windows 10 and WSL (Windows Subsystem for Linux), which gives you the ability to use Linux distros inside of Windows 10.
Although there are plenty of developers who prefer working with Windows 10, Linux and macOS users have had the upper hand since PyTorch better supports their operating systems. This could have been a driving force behind Microsoft going for more control over the library.
As of this writing, Windows users could experience several difficulties while working with PyTorch. First and foremost, they might have a hard time learning PyTorch since the insufficient test coverage of its core tutorials could cause unexpected issues. Moreover, PyPI is not officially supported for PyTorch on Windows, and thus, it might not be so easy to install this package on your system. Your experience using PyTorch on Windows could also be affected as TorchAudio, and other such important functionalities are entirely missing on the OS.
On the bright side, this move by Microsoft could bring the best out of PyTorch on Windows. Besides, the machine-learning library would also be able to perform a whole lot better on WSL, which is something that the Redmond Company has been striving for as well.
It was back in 2018 that Facebook made its widely-used library open-source. However, when this happened, it had only been a year since the social media company launched this library so that developers could build better artificial intelligence models. From this, it can be noted how much Facebook cares about spreading artificial intelligence to the masses.
If you’ve used Facebook, you might have crossed paths with its Translate feature, which deals with four dozens of most widespread languages on Facebook. Well, this feature by Facebook has been built using a deep-learning framework called Caffe2 and PyTorch. With that being said, we shouldn’t limit the significance of this library to just this feature as PyTorch has also helped with projects such as the likes of HuggingFace’s Transformers, Uber’s Pyro, and even Tesla Autopilot.
Programmers who have worked with Python would have at least heard of PyTorch, as it is one of the packages of this programming language. Courtesy of PyTorch, developers can execute heavy data science tasks that require GPU power and work with other ML-centred packages like NumPy. Plus, GitHub also shows a trend of the increasing number of projects based on PyTorch.
Conclusion
In the past, there have been various reports of Windows users facing issues working with PyTorch, which is undoubtedly one of the best machine learning libraries available. Luckily, Microsoft has come to the aid of its developers by taking charge of PyTorch to improve its support for Windows systems. Accordingly, it won’t be wrong to anticipate some significant developments shortly that will make Windows an even better OS for AI-related works.