CODEWITHSUNDEEP
amazon-web-servicesamazon-sagemakeramazon-sagemaker-studio
ezail
ezail

Reputation: 23

What is the difference of inference model between aws console and sagemaker studio model registry

When setting up an endpoint, I think the process involves creating an inference model, specifying that model in the endpoint configuration, and then using that endpoint configuration to establish the endpoint.

At that time, it appears that there are two locations for the inference model, rather than two representations of one model; they seem to exist independently:

  1. AWS console
  2. SageMaker Studio model registry

Here are my questions:

  1. In boto3, the model built by create_model module is displayed in the AWS console, while the model built by create_model_package is displayed in SageMaker Studio?
  2. The difference between the two representations is only in that if we can versioning the model?

Thanks

The docs and workshops seem only to properly describe the Python SDK usage,it would be really helpful if someone could clear this up for me

Upvotes: 1

Views: 319

Answers (1)

durga_sury
durga_sury

Reputation: 1152

You're right about the process to create the endpoint, i.e., Model -> Endpoint configuration -> Endpoint.

A model registry is basically just a registry with model versions. To deploy any of these versions from the model registry, you'll still create a Model -> Endpoint configuration -> Endpoint.

Model registry at this time is available only on the Studio UI Models, however are available on both the AWS console, and the Studio UI. In the new Studio UI, you can go to Models -> Deployable models tab and see all your models.

Upvotes: 0

Related Questions