Hugging Face Models Get One-Click Deployment to Vertex AI

Through a collaboration between Hugging Face and Google Cloud, the 'Deploy on Google Cloud' feature has launched. This allows developers to easily deploy thousands of open source models from Hugging Face Hub model cards or Google Cloud's Vertex AI Model Garden to API endpoints within their own Google Cloud account.

The key features and characteristics are as follows:

• Simple deployment process: Through the 'Deploy' menu on Hugging Face Hub, users can connect directly to the Google Cloud Console to instantly deploy models to Vertex AI or GKE (Google Kubernetes Engine).
• Vertex Model Garden integration: Using the 'Deploy From Hugging Face' option within the Google Cloud console, models can be deployed with verified hardware configurations simply by searching for the model ID.
• Performance and stability: Built on Hugging Face's production inference solution, Text Generation Inference (TGI), to support stable model serving.

With this integration, developers can reduce the burden of infrastructure management and build generative AI applications more quickly within a secure Google Cloud environment.