Fresh off a $100 million funding round, Hugging Face, which provides hosted AI services and a community-driven portal for AI tools and data sets, today announced a new product in collaboration with Microsoft. Called Hugging Face Endpoints on Azure, Hugging Face co-founder and CEO Clément Delangue described it as a way to turn Hugging Face-developed AI models into “scalable production solutions.”
“The mission of Hugging Face is to democratize good machine learning,” Delangue said in a press release. “We’re striving to help every developer and organization build high-quality, machine learning-powered applications that have a positive impact on society and businesses. With Hugging Face Endpoints, we’ve made it simpler than ever to deploy state-of-the-art models, and we can’t wait to see what Azure customers will build with them.”
The demand for AI remains high. According to a recent McKinsey survey, nearly two-thirds of companies plan to increase their investments in AI over the next two years. But implementing AI from scratch can be challenging. Moreover, many companies have strict performance, security, compliance and privacy requirements that require hosting models on tightly controlled infrastructure.
Hugging Face Endpoints is Hugging Face’s solution to the problem.
Available through Azure Machine Learning Services, Hugging Face Endpoints allows customers to tap Hugging Face models with a few clicks or lines of Microsoft Azure SDK code. After selecting a model and a task type, customers can deploy the model wherever they choose on internal infrastructure, whether for an app, website or backend service.
“Transformer models have forever changed how companies build technology with machine learning,” Jeff Boudier, product director at Hugging Face, told TechCrunch via email. “Endpoints are going to make Transformers … easily accessible to Azure customers. As of today, Hugging Face Endpoints supports all Transformers natural language processing tasks. We’re going to add support for speech and computer vision tasks, including automatic speech recognition and image classification, over the next couple of weeks.”
Hugging Face’s models focus on text analysis — specifically tasks like summarizing and generating text, extracting information and automatically answering questions. They’re largely based on the Transformer, the same model architecture underpinning OpenAI’s GPT-3 and countless other powerful AI systems.
Hugging Face Endpoints is launching in beta today — already, it’s being used by Standard Bank, a large South African bank and financial services group. It’ll be free during the preview period, with customers only having to pay for the underlying Azure compute, but the pricing model for general availability will be usage-based.
Boudier hinted that it’s only the beginning of Hugging Face’s work with Microsoft.
“This is the start of the Hugging Face and Azure collaboration we are announcing today as we work together to bring our solutions, our machine learning platform, and our models accessible and make it easy to work with on Azure. Hugging Face Endpoints on Azure is our first solution available on the Azure Marketplace, but we are working hard to bring more Hugging Face solutions to Azure,” he said. “We have recognized [the] roadblocks for deploying machine learning solutions into production and started to collaborate with Microsoft to solve the growing interest in a simple off-the-shelf solution.”