It became known about the beginning of strategic partnership between Google Cloud and Hugging Face, marking a new era in the development and use of artificial intelligence. This cooperation is aimed at democratizing advanced technologies.
During the partnership, developers will be able to utilize the power of Google Cloud infrastructure for all Hugging Face services. This will establish Google Cloud as the preferred platform for training and working with models, ensuring process optimization at each stage.
Google Cloud infrastructure, including computing power, tensor (TPU) and graphic (GPU) processors, will be leveraged to train models and create cutting-edge applications based on generative AI.
The key aspects of this cooperation are as follows:
Integration with Vertex AI, allowing for model training and configuration on the Hugging Face platform with just a few clicks, as well as accessing MLOps services from Google Cloud.
Support for deployment in Google Kubernetes Engine (GKE), giving developers the freedom to choose infrastructure and scale models.
Access to the Cloud TPU V5E, offering improved performance and reduced latency compared to previous versions.
Implementation of the A3 VM based on the NVIDIA H100 Tensor Core GPU, promising significant acceleration of learning processes and improved network throughput.
Simplification of management and billing through Google Cloud Marketplace for Hugging Face services.
“From the very beginning of the development of Transformers agents and prior work on T5 and Vision Transformer, Google has always been at the forefront of AI development and supported the open science movement,” said Cleman Delang, CEO of Hugging Face. “Through this partnership, we will facilitate the use of the latest open models for Hugging Face users and Google Cloud customers, combined with advanced optimized infrastructure and Google Cloud tools including Vertex AI and TPU. This will significantly enhance developers’ capabilities in creating and developing their own projects.”
Vertex AI and GKE are expected to be available on the Hugging Face platform in the first half of 2024.