Vertex AI endpoint deployment

[rapidhunter]

How can I utilize the mega GPU during endpoint deployment  for vertex ai work? Are there any model for examples or other resources that I can use to better grasp this?

[ricconoel]

What kind of model are you deploying on the endpoint? Can you clarify on what you mean in this statement "utilizing the GPU during endpoint deployment"? 

[rapidhunter]

Thank you for responding. It is a simple tensorflow tabular classification model, and deployment takes about 18-20 minutes after model registration. I decided to use a mega GPU since I want to shorten this time as much as I can; perhaps this decision will shorten the time for deployment.

[ricconoel]

Thank you for clarifying. As far as I know the deployment of an endpoint is handled at the backend of GCP so it is not possible to use a GPU in order to shorten the deployment time.

[rapidhunter]

okay thanks , but when I use another account for the same , it's done within 2-3 min only !!!!