\n \n \n “. concat(self. i18n. t(‘search. voice. recognition_retry’), “\n
Fresh off a $100 million investment round, Hugging Face, which provides hosted AI facilities and a network portal for AI equipment and datasets, today announced a new product in collaboration with Microsoft. Called Hugging Face Endpoints on Azure, co-founder of Hugging Face and CEO Clément Delangue described it as a way for AI models to evolve through Hugging Face into “scalable production solutions. “
“Hugging Face’s project is to democratize smart device learning,” Delangue said in a press release. “We try to help each and every developer and organization create high-quality programs based on device learning that have a positive effect on society. “and business. With Hugging Face Endpoints, we’ve made it less difficult than ever to deploy models, and we can’t wait to see what Azure consumers will create with them. “
Demand for AI remains high. According to a recent McKinsey survey, nearly two-thirds of corporations plan to increase their AI investments over the next two years. But implementing AI from scratch can be challenging. In addition, many corporations have stringent performance, security, compliance, and privacy requirements that require models in a tightly controlled infrastructure.
Hugging Face Endpoints is Hugging’s problem.
Available through Azure Machine Learning Services, Hugging Face Endpoints allows consumers to take advantage of Hugging Face models with just a few clicks or lines of Microsoft Azure SDK code. After settling on a template and a task type, consumers can deploy the template anywhere they need for the internal environment. infrastructure, whether for an application, website, or back-end service.
“Transformer models have replaced the way corporations build generation with device learning,” Jeff Boudier, product manager for Hugging Face, told TechCrunch via email. “The endpoints will make Transformers. . . is readily available to Azure customers. Starting today, Hugging Face Endpoints helps in all Transformers herbal language processing tasks. We will upload help for speech popularity and computer vision tasks, adding automatic speech popularity and symbol classification, over the next two weeks. “
Hugging Face templates concentrate on text research, especially on responsibilities such as summarizing and generating text, extracting information, and answering questions automatically. They rely largely on the Transformer, the same style of architecture that underpins OpenAI’s GPT-3 and many other resilient AIs. Systems
Hugging Face Endpoints will launch in beta today; it is already being used through Standard Bank, a leading banking and monetary services group in South Africa. You’ll lose the preview period, and consumers will only have to pay for the underlying Azure compute, but the Pricing style for general availability will be based on usage.
Boudier hinted that this is just the beginning of Hugging Face’s work with Microsoft.
“This is the beginning of the Hugging Face and Azure collaboration we announced today as we work together to make our responses, device learning platform, and models available and make it easier to work with Azure. Hugging Face Endpoints on Azure is our first solution available in the Azure Marketplace, but we are striving to bring more Hugging Face responses to Azure,” he said. “We identified [the] barriers to implementing device learning responses in production and began working with Microsoft to address the growing interest in a simple, out-of-the-box solution. “