{"id":617,"date":"2025-05-19T09:00:28","date_gmt":"2025-05-19T16:00:28","guid":{"rendered":"https:\/\/devblogs.microsoft.com\/foundry\/?p=617"},"modified":"2025-05-19T06:05:07","modified_gmt":"2025-05-19T13:05:07","slug":"microsoft-and-hugging-face-expand-partnership-to-accelerate-open-source-ai-innovation-on-azure-ai-foundry","status":"publish","type":"post","link":"https:\/\/devblogs.microsoft.com\/foundry\/microsoft-and-hugging-face-expand-partnership-to-accelerate-open-source-ai-innovation-on-azure-ai-foundry\/","title":{"rendered":"Microsoft and Hugging Face expand collaboration to accelerate Open-Source AI Innovation on Azure AI Foundry"},"content":{"rendered":"

Today, Microsoft and Hugging Face are excited to announce an expanded collaboration that puts over ten thousand Hugging Face models at the fingertips of Azure developers.<\/span><\/b>\u00a0<\/span>\u00a0<\/span><\/p>\n

Azure AI Foundry Models now allows immediate deployment of the most popular open models on Hugging Face, spanning <\/span>text, vision, speech, and multimodal<\/span><\/b> models.<\/span> This expanded collaboration advances our shared goal to democratize AI by making it easier for developers and enterprises to access, deploy, and scale the latest open models on Azure. <\/span>This also makes Azure AI Foundry as the platform that offers one of the largest and most diverse open-source model collections on a hyper scale cloud platform.<\/span>\u00a0<\/span><\/p>\n

\u201cThis expanded collaboration <\/span>is a reflection of<\/span> our deep commitment to open-source AI. Hugging Face has emerged as the <\/span>central h<\/span>ub<\/span> hosting<\/span> millions of open models and serving as the default launchpad for open-source AI innovation.”\u00a0 said Asha Sharma, Corporate Vice President at Microsoft<\/strong>. \u201cBy combining Hugging Face\u2019s vibrant developer ecosystem with Azure\u2019s enterprise-grade infrastructure, we\u2019re enabling customers to innovate faster and more securely with the best models the community has to offer.\u201d<\/span><\/span><\/p><\/blockquote>\n

Hugging Face<\/span> is home to millions of open <\/span>models.<\/span> O<\/span>rganizations<\/span><\/span>,<\/span><\/span><\/span> large <\/span>and<\/span> small<\/span>,<\/span> along with <\/span>a large community of <\/span>8 million<\/span> AI <\/span>builders<\/span><\/span>,<\/span><\/span><\/span> use<\/span> the<\/span> Hugging Face<\/span> platform to <\/span><\/span>access and share models, <\/span>datasets<\/span> and applications<\/span><\/span>.<\/span> They release model weights <\/span>and<\/span> cards <\/span>for their models <\/span>on Hugging Face <\/span>organizations<\/span>, <\/span>contribute to Hugging Face <\/span>libraries<\/span> such as transformers and diffusers to <\/span>enable<\/span> new model architectures<\/span>,<\/span> and <\/span>demonstrate<\/span> model <\/span>performance<\/span> using<\/span> evaluation<\/span> metrics on Hugging Face open model leaderboards. <\/span>Leveraging<\/span> all this innovation <\/span>in an enterprise environment, <\/span>where you need to s<\/span>cale<\/span> throughput, optimize latency, <\/span>screen for malicious <\/span>code<\/span>,<\/span> a<\/span>nd <\/span>secure <\/span>endpoints <\/span>with private networks <\/span>can be<\/span> a challenge. <\/span>The goal of this expanded collaboration is to <\/span>bring<\/span> open<\/span>–<\/span>source<\/span> AI innovation <\/span>to Azure<\/span> through a <\/span>secure, <\/span>scalable<\/span> and reliable <\/span>integration between <\/span>the <\/span>Hugging Face h<\/span>ub<\/span> and <\/span>Azure AI Foundry<\/span>.<\/span><\/span><\/p>\n

Expanding Modalities<\/h3>\n

The refreshed integration enables<\/span> Azure AI<\/span> developers to incorporate <\/span>more modalities in their applications<\/span> and<\/span> facilitate<\/span> a <\/span>variety<\/span> of <\/span>tasks<\/span> including<\/span> text generation, feature extraction, fill<\/span>–<\/span>mask, translation, <\/span>identifying <\/span>sentence similarity, <\/span>image classification, image segmentation, text to image<\/span> generation<\/span>, image to text<\/span> conversion<\/span>, automatic speech recognition<\/span> and <\/span>audio-classification<\/span>. <\/span>Whether <\/span>you’re<\/span> starting<\/span> foundational research or deploying AI at scale, you <\/span>now<\/span> have access to one of the most robust open-source AI ecosystems<\/span>, <\/span>backed by the enterprise readiness of Microsoft Azure.<\/span><\/span><\/p>\n

\"\"<\/a><\/p>\n

Microsoft is also working with Hugging Face to bring in additional models enabling tasks like text to speech, audio to audio, image to 3D and text to video. Support for trending models of Hugging Face hub on Foundry across these new modalities is coming soon, expanding what’s possible for developers to build multimodal AI solutions on Azure.<\/p>\n

As Clem Delangue, co-founder and CEO of Hugging Face<\/strong>, puts it, \u201cWe\u2019re enabling companies to take control of their AI destiny, deploying the best open models securely within their Azure account, to build AI applications they can trust and verify.\u201d<\/p><\/blockquote>\n

\u00a0<\/span><\/span>Top Providers, Trending Models\u2014Available Day One<\/span><\/span><\/h3>\n

Through this expanded collaboration, we<\/span> will<\/span> also <\/span>offer the latest models from top <\/span>model<\/span> providers in the <\/span>Azure AI <\/span>Foundry catalog<\/span> on the same day <\/span>they’re<\/span> released on the Hugging Face hub<\/span>, providing<\/span> day one <\/span>access to the <\/span>latest<\/span> models <\/span>for one-click deployment on Foundry.<\/span><\/span><\/p>\n

We will<\/span> also <\/span>bring<\/span> in customized variants of trending <\/span>open<\/span> models, giving enterprises and developers access to fine-tuned, production-ready <\/span>models <\/span>meeting functional and performance needs.<\/span><\/span><\/p>\n

Security First<\/strong><\/h3>\n

All models made available in the Hugging Face collection in the Azure AI Foundry catalog for discovery and deployment have been scanned by Hugging Face to be safe from vulnerabilities. As part of this expanded collaboration, Microsoft will work with Hugging Face to also host certain model weights on Azure, that will enable customers to deploy these models as secure endpoints with no external network egress to Hugging Face hub.<\/p>\n

Beyond Models: Tools, Agents, and More<\/strong><\/h3>\n

In the future, we will also work to bring community driven apps from Hugging Face, expanding the catalog to containerized tools to power agents with the smolagents framework. This will enable developers to build modular agentic applications using open and composable components, all within a scalable, production-grade cloud environment.<\/p>\n

Consume using Azure AI Inference SDK<\/strong><\/h4>\n

Models offered with Text Generation Inference (TGI) and Text Embedding Inference (TEI) offer an OpenAI Chat Completion API compatible layer under the API routes\u00a0\/v1\/chat\/completions and \/v1\/embeddings, check the full specification in the\u00a0OpenAI Chat Completion Create Documentation<\/a>. These models can be consumed using Azure AI Inference SDK<\/a>.<\/p>\n

Python Code sample:<\/strong><\/p>\n

Pre-requisite: Install the Azure AI Inferencing package using pip install azure.ai.inference<\/span><\/p>\n

Add your Azure AI Foundry endpoint and API Key to the following sample code sample and run inference:<\/p>\n

import os\r\nfrom azure.ai.inference import ChatCompletionsClient\r\nfrom azure.core.credentials import AzureKeyCredential\r\nfrom azure.ai.inference.models import SystemMessage, UserMessage\r\n\r\nendpoint = \"https:\/\/<endpoint-name>.<region>.inference.ml.azure.com\/v1\"\r\nkey = os.getenv(\"AZURE_AI_CHAT_KEY\", \"key here\")\r\n\r\nclient = ChatCompletionsClient(\r\n    endpoint=endpoint,\r\n    credential=AzureKeyCredential(key),\r\n)\r\n\r\nresponse = client.complete(\r\n    stream=True,\r\n    messages=[\r\n        SystemMessage(\"You are a helpful assistant.\"),\r\n        UserMessage(\"I want to go to Iceland, which is the best time of the year to visit?\"),\r\n    ],\r\n)\r\n\r\nfor update in response:\r\n    if update.choices and update.choices[0].delta:\r\n        print(update.choices[0].delta.content or \"\", end=\"\", flush=True)\r\n    if update.usage:\r\n        print(f\"\\n\\nToken usage: {update.usage}\")\r\n\r\nclient.close()<\/code><\/pre>\n
Learn More and Get Started<\/h3>\n
As the open-source AI landscape continues to evolve, our expanded collaboration with Hugging Face will remain focused on bringing the latest in AI innovation to Azure AI developers.<\/p>\n
\ud83d\udc49 Start exploring the Azure AI + Hugging Face collection: https:\/\/aka.ms\/FoundryHuggingFaceCollection<\/a><\/p>\n
\ud83d\udc49 Join us at Microsoft Build to see it in action<\/p>\n","protected":false},"excerpt":{"rendered":"
Microsoft and Hugging Face announce an expanded collaboration that puts over ten thousand Hugging Face models at the fingertips of Azure developers.<\/p>\n","protected":false},"author":190443,"featured_media":662,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[1],"tags":[4,34,2],"class_list":["post-617","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-microsoft-foundry","tag-generative-ai","tag-microsoft-build","tag-microsoft-foundry"],"acf":[],"blog_post_summary":"
Microsoft and Hugging Face announce an expanded collaboration that puts over ten thousand Hugging Face models at the fingertips of Azure developers.<\/p>\n","_links":{"self":[{"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/posts\/617","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/users\/190443"}],"replies":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/comments?post=617"}],"version-history":[{"count":0,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/posts\/617\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/media\/662"}],"wp:attachment":[{"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/media?parent=617"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/categories?post=617"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/tags?post=617"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}