|Google’s Hugging Face deal puts ‘supercomputer’ power behind open-source AI
In a groundbreaking move, Google Cloud has joined forces with Hugging Face, transforming the landscape for AI developers seeking unprecedented power in open-source artificial intelligence. This strategic partnership not only eliminates the need for a Google Cloud subscription but also provides developers on the Hugging Face platform with cost-effective access to Google's formidable cloud infrastructure, boasting thousands of GPUs and the coveted tensor processing units (TPUs), including Nvidia's sought-after and export-restricted H100s.
Elevating AI Development with Hugging Face
Access to Google's Supercomputing Resources
With this collaboration, AI developers utilizing Hugging Face gain access to Google's cloud setup, leveraging the muscle of thousands of GPUs seamlessly integrated with services like Vertex AI. The implications are profound, as developers can now build, train, and deploy AI models without the financial constraints of a traditional Google Cloud subscription.
Hugging Face: A Hub of AI Innovation
Hugging Face stands out as one of the most popular AI model repositories, housing a plethora of open-sourced foundational models such as Meta's Llama 2 and Stability AI's Stable Diffusion. Beyond this, it serves as a repository for various databases essential for model training. Currently hosting over 350,000 models, Hugging Face operates similarly to GitHub, allowing developers to collaborate, share, and upload their own models.
Valuation and Support
The significance of Hugging Face in the AI landscape is underscored by its impressive valuation of $4.5 billion. In the past year alone, it has garnered substantial support, raising $235 million from industry giants like Google, Amazon, and Nvidia. This influx of capital positions Hugging Face as a key player driving innovation in the AI space.
The Future of AI Development: What to Expect
Expanding Capabilities with Vertex AI
Google has announced that Hugging Face users can soon leverage Vertex AI, a powerful AI app-building platform. Additionally, developers can harness the capabilities of the Kubernetes engine for training and fine-tuning models, with these features expected to be available in the first half of 2024. This development further cements Google's commitment to nurturing the open-source AI ecosystem.
Google Cloud's Stance on Open-Source AI
In a statement, Google emphasized that its collaboration with Hugging Face is a testament to Google Cloud's unwavering support for the development of the open-source AI ecosystem. While some of Google's models are already present on Hugging Face, it's important to note that certain flagship models, like Gemini powering the chatbot Bard and the text-to-image model Imagen, remain proprietary and closed source.
In conclusion, the Google-Hugging Face partnership marks a significant stride towards democratizing AI development. With cost-effective access to Google's supercomputing resources, developers on Hugging Face can unlock new possibilities, fostering innovation and collaboration in the ever-evolving landscape of artificial intelligence.