Published on November 16, 2023, 4:57 pm
Microsoft is making a big push with generative AI experiences, and it’s aiming to do so not only on the cloud but also on Windows devices. The company recently unveiled Windows AI Studio at its Ignite 2023 conference, with plans to launch it in preview in the coming weeks. This new platform serves as a successor to the now-defunct AI Platform for Windows developers and brings together AI tools and a catalog of generative AI models that developers can customize, fine-tune, and deploy for offline use in their Windows apps.
What sets Windows AI Studio apart is its flexibility. It doesn’t force developers to deploy models either offline or in the cloud. Instead, it provides them with options to run models locally or remotely in data centers—a hybrid local-cloud configuration if needed. The goal is to streamline AI development on Windows by equipping developers with the necessary tools and resources as quickly as possible.
One standout feature of Windows AI Studio is its model catalog, which includes models from Microsoft himself as well as third-party repositories like Hugging Face’s. With this catalog, developers have the opportunity to select, configure, fine-tune, and test their models using their own data sets before deciding where to deploy them—locally or in the cloud.
The lack of a guided step-by-step interface has been a hurdle for many developers looking to fine-tune their models for deployment into generative AI applications. However, Windows AI Studio solves this problem by offering an intuitive interface that allows developers to focus on coding while reducing the heavy lifting required for deploying these features into their apps.
Windows AI Studio will soon be upgraded with an exciting feature called prompt flow. This functionality enables apps to automatically switch between smaller, faster models running locally on machines and more powerful cloud-hosted models when necessary. In addition, supported hardware with dedicated AI accelerator chips or GPUs will open up opportunities for even more capable local models—notably including Meta’s text-generating Llama 2, thanks to Microsoft’s expanded partnership with Meta, and Stability AI’s text-to-image model Stable Diffusion XL.
To make development even more accessible, Microsoft plans to release Windows AI Studio as an extension for VS Code, an open-source code editor. This move will provide developers with greater flexibility in choosing whether to run their models on the cloud or locally on Windows.
Developers facing challenges in securing sufficient cloud resources for running AI models at scale will find Windows AI Studio’s local deployment capabilities particularly valuable. However, the ease of running performant models locally remains to be seen until the public preview is available.
With the upcoming launch of Windows AI Studio, Microsoft is offering developers a comprehensive platform for generative AI development on Windows devices. By combining cutting-edge tools with a rich model catalog and support for local and hybrid deployments, this new offering aims to accelerate progress in the world of generative AI applications.