Microsoft’s hands on Hugging Face, the number one in machine learning
Hugging Face is considered the world's top independent AI model checker and has been valued at $2 billion. Here's what it does with Microsoft and why it's key to Bill Gates' company in the AI competition
Microsoft's race to be first in the global artificial intelligence (AI) competition seems unstoppable . After the 10 billion dollar investment in OpenAI – the software house that created ChatGpt -, Bill Gates' company has set its sights on Hugging Face, a machine learning platform, recently valued at 2 billion dollars and meeting of AI masterminds.
WHAT MAKES HUGGING FACE
“We're on a mission to democratize good machine learning, one endeavor at a time,” says Hugging Face. And in fact, its goal is to democratize artificial intelligence through open source and open science, giving developers the opportunity to discuss anything on the platform.
Even its symbol, the emoji representing a hug, is meant to show openness and inclusiveness, as well as appreciation, support and attention.
As Quartz points out, Hugging Face is famous for its vast library, Transformers, which allows you to build solutions based on the AI subsets of natural language processing, machine learning, and computer vision.
HUGGING FACE NUMBERS
Founded in 2016 in New York, it is now considered the world's top independent AI model checker, according to Business Wire .
The company closed a $100 million financing round last year that raised its valuation to $2 billion. At the time, according to Supervised, Hugging Face had annual revenues of between $7 million and $10 million.
Today, it hosts over 200,000 open-source templates.
THE INTEREST OF MICROSOFT
Microsoft, which has sensed the company's potential, is taking steps to ensure it has a place within its products, "even if they seem embryonic and small for now," writes Supervised .
The closest collaboration dates back to last year, when Hugging Face released an endpoint for Azure, Microsoft's cloud platform, in May. But now it has a real role in the Azure AI workflow and Supervised wonders if this will change the original grand ambitions.
THE COLLABORATION BETWEEN MICROSOFT AND HUGGING FACE
In a March release reported by Quartz , Microsoft announced a partnership with Hugging Face to bring open-source models to the Azure Machine Learning platform.
At the moment, according to the article, Microsoft already has 500 team members on the platform and has shared 246 models – programs that use training data to recognize patterns and make decisions – that have been downloaded millions of times by users.
“It is a milestone in our mission to accelerate AI initiatives and bring innovative solutions to market quickly and securely,” said John Montgomery, vice president of Azure AI Platform, explaining that the collaboration will simplify the process of deploying and scaling AI. large language models.
However, Microsoft has put a damper on the excitement by warning users that while Hugging Face's Transformer models are easy to test on the platform, they "can be difficult to implement and scale for production-grade inference endpoints."
This is a machine translation from Italian language of a post published on Start Magazine at the URL https://www.startmag.it/innovazione/le-mani-di-microsoft-su-hugging-face-il-numero-uno-del-machine-learning/ on Sun, 04 Jun 2023 15:30:49 +0000.