Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More.
As businesses across various industries strive to implement their AI visions, vendors are stepping up to provide comprehensive resources. A prime example is the new collaboration between Google and Hugging Face, aimed at simplifying access to Google Cloud services and speeding up the development of open generative AI applications.
With this partnership, teams using Hugging Face’s open-source models can now train and deploy them on Google Cloud. This offers them full access to Google’s AI capabilities, including Vertex AI, tensor processing units (TPUs), and graphics processing units (GPUs).
Google has consistently led AI advancements, from pioneering the Transformers paper to developing models like T5 and the Vision Transformer. This partnership will allow Hugging Face users and Google Cloud customers to use the latest open models alongside Google’s top-notch AI infrastructure and tools, significantly enhancing developers’ ability to create their own AI models.
What does this mean for Hugging Face users? Hugging Face has become a central hub for AI, housing over 500,000 AI models and 250,000 datasets. More than 50,000 organizations use the platform for their AI projects. Concurrently, Google Cloud has been expanding its AI-focused infrastructure and tools while supporting open AI research.
Through this partnership, Hugging Face users on Google Cloud can train, fine-tune, and deploy their models using Vertex AI, Google’s comprehensive MLOps platform, to create new generative AI applications. This integration will be accessible with just a few clicks from the Hugging Face platform, allowing model training and deployment on Google Kubernetes Engine (GKE). This setup provides a flexible infrastructure to scale models using Hugging Face’s deep learning containers on GKE.
Additionally, developers will benefit from Google Cloud’s advanced hardware, including TPU v5e, A3 VMs with Nvidia H100 Tensor Core GPUs, and C3 VMs with Intel Sapphire Rapid CPUs.
Models will be easily deployable for production on Google Cloud using inference endpoints. AI developers can boost their applications with TPU on Hugging Face spaces. Organizations can manage their Google Cloud account to streamline the usage and billing of their Enterprise Hub subscription, as explained by Jeff Boudier and Philipp Schmid from Hugging Face.
It’s important to note that while this collaboration has been announced, the new features, including Vertex AI and GKE deployment options, will become available to Hugging Face Hub users in the first half of 2024.