Executive Summary - Microsoft and Meta Partner
Meta and Microsoft as partners have moved closer together:
- The first move Meta Hosting its services on Azure Cloud
- The second their immersive partnership on Meta Quest 3 & Microsoft products, Microsoft Mesh and the 3D meeting product being key.
- The third the LLama 2 Large Language models support that both announced, that transform the Ai world.
The three changes have strengthened significantly Meta's enterprise credentials for Meta for Work.
Details of the Meta and Microsoft Llama 2 - AI Support partnership
In recent months, advancements in AI innovation have sparked a wave of transformative possibilities. At Microsoft Inspire, Meta and Microsoft announced support for the Llama 2 family of large language models (LLMs) on Azure and Windows.
Meta and Microsoft share a commitment for AI's development and its benefits. This partnership strengthens the potential for collaboration and innovation within immersive spaces, fostering integration within Microsoft Teams and the broader meta verse.
Now Azure customers can easily and safely fine-tune and deploy the 7B, 13B, and 70B-parameter Llama 2 models on Azure, the platform that supports the most widely adopted frontier and open models. In addition, Llama 2 will be optimised to run locally on Windows, allowing Windows developers to target the DirectML execution provider through the ONNX Runtime.
The collaboration between Meta and Microsoft is set to transform the future of virtual reality by enhancing Microsoft Mesh and Mesh for Teams, enabling immersive experiences and custom connections. By leveraging these advanced capabilities, developers can create tailored virtual reality experiences that foster deeper connections and collaboration, redefining the potential of the metaverse and immersive spaces.
Microsoft's growing partnership with Meta enhancing its Enterprise credentials
Meta and Microsoft have been longtime partners on AI, starting with a collaboration to integrate ONNX Runtime with PyTorch to create a great developer experience for PyTorch on Azure, and Meta’s choice of Azure as a strategic cloud provider. They have announced that they plan to build on their partnership to accelerate innovation in the era of AI.
Azure’s AI platform is uniquely designed from the ground up, incorporating facility, hardware, and software optimisations to support the world’s leading AI organisations in building, training, and deploying some of the most demanding AI workloads. With the availability of the Llama 2 models on Azure AI, developers can leverage Azure AI’s powerful tools for model training, fine-tuning, inference, and enhanced AI safety capabilities.
This integration also extends to Meta and Microsoft Mesh, empowering businesses and organisations to create more engaging connections. With Mesh for Teams, the future of virtual reality and virtual meetings becomes a reality, as VR devices and cutting-edge technology enable a new level of presence in business interactions.
The inclusion of the Llama 2 models in Windows strengthens its position as the ideal platform for developers to build AI experiences tailored to their customers’ needs. By utilising world-class tools such as Windows Subsystem for Linux (WSL), Windows Terminal, Microsoft Visual Studio, and VS Code, developers can unlock new possibilities in creating advanced AI applications.
This advancement also supports the integration of Meta and Microsoft Teams, enhancing connection through custom experiences. With a focus on the future of VR, Mesh enables the creation of immersive virtual workspaces for enterprise environments. Hybrid meetings are transformed as colleagues interact using avatars, fostering immersive employee experiences in Mesh environments, increases productivity and engagement. Read more about how productivity is affected.
Expanding Azure AI model catalog and Windows availability
Llama 2 is the latest addition to the growing Azure AI model catalog. Currently in public preview, this model catalog serves as a hub for foundation models, empowering developers and machine learning (ML) professionals to easily discover, evaluate, customise, and deploy pre-built large AI models at scale.
The catalog eliminates the need for users to manage infrastructure dependencies when operationalising Llama 2. It provides turnkey support for model fine-tuning and evaluation, using powerful optimisation techniques such as DeepSpeed and ONNX Runtime to significantly enhance the speed of model fine-tuning.
Windows developers will be able to easily build new experiences using Llama 2, accessible via the GitHub repository. With Windows Subsystem for Linux and highly capable GPUs, developers can fine-tune large language models (LLMs) to meet their specific needs directly on their Windows PCs.
This capability enhances the connection with custom experiences and virtual collaborations, allowing Teams users and clients to create more dynamic and personalised sessions. By integrating these developments with Meta and Microsoft Mesh, developers can improve interactions within virtual environments and team socials. The Microsoft Community Hub serves as a platform for sharing insights and engaging with these innovations through blogs and discussions. This advancement opens up new possibilities for virtual experiences and collaborations across various platforms.
Building responsibly with Azure
Responsible AI is at the heart of Microsoft's approach to AI. For years, they have invested heavily in making Azure the place for responsible, cutting-edge AI innovation, whether customers are building their own models or using pre-built and customisable models from Microsoft, Meta or OpenAI.
At Microsoft, they mitigate potential risks presented by large language models through an iterative, layered approach that includes experimentation and measurement. Azure AI customers can test Llama 2 with their own sample data to see how it performs for their specific use. They can then use prompt engineering and retrieval augmented generation (RAG) techniques to develop, evaluate, and optimise meta-prompts for their applications, delivering safer and more reliable experiences for end users.
Services like Azure AI Content Safety add another layer of protection, helping ensure a safer online experience with AI apps. Part of Microsoft's collaboration with Meta led to the integration of Meta’s safety techniques with Azure AI Content Safety, ensuring that by default, Llama 2 model deployments in Azure AI come with a layered safety approach.
The expansion of the model catalog with Llama 2 and the partnership with Meta is a significant step forward in achieving a responsible, open approach to AI. This collaboration extends into the virtual world, enhancing the functionality of Microsoft Mesh for companies.
