AWS And Hugging Face Partner To Enable 100,000+ AI Models Using Amazon Chips

As AI companies and startups strive to push the growth of artificial intelligence (AI) and generative artificial intelligence (GenAI), they find themselves enabling the availability of tools, hardware and open-source models through cost-effective platforms, allowing smaller players and developers to bloom in the sector.

Hugging Face is one such platform that helps users build, deploy and train machine learning models, by providing them with the required infrastructure to test, run and deploy AI technology in products and live applications.

As per its last funding round in August 2023, the AI startup was valued at $4.5 billion and is backed by leaders such as Amazon, Google, NVIDIA, Salesforce and more.

Today, the platform has become a popular destination for AI researchers and developers to share chatbots, virtual assistants and other AI-powered software.

However, to keep up with industry capabilities, such platforms also need to provide its users with powerful infrastructure and that’s what Hugging Face is looking to fulfill with its latest announcement.

So, what did the widely popular AI startup reveal? Let’s explore!
 

What Did AWS And Hugging Face Announce?

 

• Through a blog post published on its website, Hugging Face announced that it was bringing Amazon’s high performance machine learning (ML) chip widely available to Hugging Face Hub users.

• As per the release, Hugging Face has been working with AWS’s product and engineering teams to bring the performance and cost-efficiency of AWS Trainium and Inferentia chips for its uses.

• The AI startup comes with an open-source library that simplifies the training and deployment of models on such accelerators.

• Through the move, Hugging Face will enable over 100,000 public models, 14 new model architectures and 6 new machine learning tasks on AWS Inferentia2 through Amazon SageMaker.

• Hugging Face even introduced a new way to deploy Large Language Models or LLMs on SageMaker using Inferentia/Trainium chips for models such as Meta Llama 3.

• Now, users “can deploy a Llama3 model on Inferentia2 instances on SageMaker to serve inference at scale and benefit from SageMaker’s complete set of fully managed features for building and fine-tuning models, MLOps and governance.”

• Furthermore, once developers tweak an open-source AI model, they can use the platform to power their software.

• This is where AWS also hopes to draw AI developers to their cloud services to deliver their software.

• Ahead of this, the startup’s Hugging Face Inference Endpoints will also bring support for AWS Inferentia2, which will allow users to deploy models they’re interested in just a few clicks on Inferentia2.

• The startup’s future plan includes expanding the scope of models enabled for deployment through Hugging Face Inference Endpoints, followed by adding support for Diffusion and Embedding models.

 

What Did Stakeholders Say?

 

• Jeff Boudier, head of product and growth at Hugging Face, said, “One thing that's very important to us is efficiency - making sure that as many people as possible can run models and that they can run them in the most cost effective way.”

• Matt Wood, the VP of artificial intelligence products at AWS, said, “You train these models maybe once a month. But you may be running inference against them tens of thousands of times an hour. That's where Inferentia2 really shines.”

Do you think this partnership will help AWS build a better market position and boost its market share in the AI sector? Do you think it will also benefit Hugging Face users?

Let us know in the comments below!