Hugging Face and AWS expand cloud partnership

Hugging Face Inc., the operator of a preferred platform for internet hosting machine studying fashions, is partnering with Amazon Net Providers Inc. to streamline synthetic intelligence growth tasks. 

The businesses announced the partnership at present. It expands upon an present collaboration that Hugging Face and AWS launched in early 2021.

“Generative AI has the potential to rework whole industries, however its value and the required experience places the expertise out of attain for all however a choose few firms,” mentioned AWS Chief Government Officer Adam Selipsky. “Hugging Face and AWS are making it simpler for patrons to entry fashionable machine studying fashions to create their very own generative AI functions with the best efficiency and lowest prices.”

New York-based Hugging Face is backed by greater than $160 million in funding. It operates a platform just like GitHub that builders use to host open-source AI fashions, in addition to associated technical belongings corresponding to coaching datasets. The platform shops the code for greater than 100,000 neural networks.

As a part of the expanded partnership introduced at present, Hugging Face will use AWS as its most well-liked public cloud. Moreover, the startup is rolling out a brand new integration with the Amazon SageMaker machine studying platform. The platform consists of greater than a half dozen cloud providers that builders can use to construct, practice and deploy AI fashions.

The newly introduced integration will allow builders to deploy neural networks hosted by Hugging Face on SageMaker in a number of clicks. After an AI mannequin is uploaded to SageMaker, it may be educated utilizing cloud situations powered by AWS Trainium chips. The chips are particularly optimized for AI coaching duties. 

Neural networks deployed from Hugging Face to AWS additionally work with different forms of cloud situations, together with these powered by the AWS Inferentia accelerator collection. Inferentia accelerators are chips optimized to carry out inference, or the duty of working AI fashions in manufacturing after the coaching part is full.

“The way forward for AI is right here, however it’s not evenly distributed,” mentioned Hugging Face CEO Clement Delangue. “Amazon SageMaker and AWS-designed chips will allow our group and the bigger machine studying group to transform the most recent analysis into brazenly reproducible fashions that anybody can construct on.”

The brand new integration will complement the Hugging Face AWS Deep Studying Containers that the businesses already provide to builders as a part of their partnership. The containers make AI fashions from Hugging Face out there in a prepackaged format that’s simpler to deploy in public cloud environments.

Picture: AWS