Emerge 2024 Project of the Year: Open-Source AI Platform Hugging Face

In a world being shaped by artificial intelligence, few companies have By 2024, they've left footprints like the open-source project Hugging Face.

What started as a chatbot app has since evolved into an open source AI hub, becoming an essential resource for researchers, developers and businesses alike. In the year By 2023, following multiple investment rounds, Hugging Face is valued at $4.5 billion.

Hugging Face is the 2024 Emerge Project's commitment to AI's transformative role and democratization of machine learning. With visionary leadership, open source tools, and a strong focus on ethics, it inspires researchers and startups around the world. As well as a thriving online community of open source AI enthusiasts, Hugging Face has become the standard bearer for responsible and collaborative AI innovation.

Table of Contents

Toggle

What is a face hug?

In the year Founded in 2016 by French entrepreneurs Clement Delange, Julien Chaumond and Thomas Wolff and based in New York City, EmbraceFace is an open source platform for machine learning and natural language processing.

Comprised of over one million AI models, 190,000 datasets, and 55,000 demo apps, Hugging Face enables developers, researchers, and data scientists to build, train, share, and deploy AI models.

“We started as a game company and when we started opening up some of our research code, we realized we had a really big impact. That led to the Transformers library and we saw the impact and excitement in the community,” co-founder and chief science officer Wolff told Decrypt. “Open source is democratizing machine learning.” We think it's a key way to bring it.”

At its core is the Transformer library, which provides pre-trained modern models for many functions. Users can explore models through browser-based reference widgets, access them through APIs, and deploy them across computing environments. Face Hug also encourages collaboration by allowing users to share and edit models, using a central data repository (Hub) where users can test and contribute to higher-level AI models.

In AI, optimization refers to taking a pre-trained AI model—which contains the weights and features learned from the original datasets to train the model—and adapting it to perform a specific task or improve performance on a specific data set.

“Open science and open source AI can protect blackbox systems, make companies more accountable and help [solve] Today's challenges—like reducing bias, reducing misinformation, promoting copyright, and rewarding all stakeholders in the value creation process—artists and content creators,” said co-founder and CEO DeLange on X (formerly Twitter).

Democratic AI

A common refrain in a decentralized and open-source society is to “democratize AI,” or enable individuals to use AI for social good, innovation, and solving complex problems outside of the control of corporations and governments.

In an industry dominated by proprietary technology and closed ecosystems, Hugging Face stands out for providing the global AI community with cutting-edge tools for free. DeLonge reiterated Hugging Face's commitment to democratizing AI at a June 2023 Congressional hearing of the Science, Space, and Technology Committee.

“Hugging Face is a community-focused company based in the US with a mission to democratize good machine learning,” DeLange said during the hearing. “Primarily, our mission is to create a platform for open source and open science, a platform to host machine learning models and datasets, and an infrastructure to support research and resources to help people of all backgrounds contribute to AI.

Democratic AI especially affects underrepresented regions and industries, where researchers and small startups often don't have the resources to compete with tech giants.

“The long-standing and widening division of resources, particularly between industry and academia, limits who can contribute to innovative research and applications,” DeLange told Congress. “We're leveraging US national AI research resources and resources for small businesses and startups doing public interest research.”

Cooperation over competition

Emphasizing Hug Face's collaborative spirit, the company has worked with other big names in AI, including Google, AWS, Meta, Nvidia and Microsoft.

In January, Hugging Face partnered with Google Cloud to combine its own open models with Google's infrastructure, all with the goal of making AI more accessible. That same month, Hugging Face introduced the company's inaugural Hallucination Leaderboard.

“The challenge now is to have enough startups and teams ready to deploy models across multiple verticals,” Wolff said. “There's no need to wait for GPT-5; now is the time to build AI applications by learning how to use, evaluate and adapt these models in today's world.”

In May, HugFace expanded its partnership with Microsoft by 2022, giving developers a broader infrastructure and tools to create more powerful versions of their Copilot AI models. Later that month, Amazon announced a new partnership with Hugging Face to make it easier for developers to run AI models using Amazon's computer chips.

Computer chip giant Nvidia has announced a partnership with Hugging Face that brings its NVIDIA Accelerated Demand Services to an open source platform, allowing developers to deploy AI models like Llama 3 with up to five times faster simulation.

In October, Hugging Face launched HuggingChat, the platform's answer to OpenAI's ChatGPT. HuggingChat allows users to choose from a variety of open source AI models for its text generation capabilities. Following that, Hugging Face Generative AI Services, or HUGS, was released, allowing developers to deploy and train AI models offline in a personalized environment.

At the Robotics Education Conference in Germany in November, Hugging Face and Nvidia announced a partnership to push open-source robotics forward, combining Hugging Face's robotics platform LeRobot with NVIDIA's AI tools for simulation and real-world training — all with the goal of making robots smarter and more efficient.

But it wasn't always a soft face to hug. In November, the company faced backlash after it was revealed that a dataset of more than a million posts was created using content deleted from the growing Bluesky social media platform, before being removed the next day.

“I have removed bluesky data from the repo. While I would like to support tool development for the platform, I recognize that this approach violates the principles of transparency and consent in data collection,” Hugging Face Machine Learning librarian Daniel Van Streen wrote in Bluesky. “I apologize for this mistake.”

The future of hugging

Moving forward to 2025, Hugging Face CEOs lay out their predictions for the next year in AI – including the first major public backlash related to AI, a large company's market capitalization cut in half due to AI, and more than 100,000 personal AI robots on the rise. Order.

“With 15 million AI builders embracing, we're starting to see AI's potential for economic and job growth,” DeLange tweeted.

Wolff shares similar optimism about open-source AI and the future of robotics moving to 2025, with more energy-efficient models, open-

“There are many things that excite me about the future, just to name a few,” Wolff said. “Smaller models that can be much more energy efficient, the addition of open source robotics, and the extension of all the tools we've got with AI into science, like weather forecasting and material discovery.”

Embrace Face has played a major role in the evolution of AI in 2024, driving innovation, global access and transparency while lowering barriers for startups and developers to create multiple AI solutions.