TimeSeriesEndpoint

class autogluon.cloud.TimeSeriesEndpoint(endpoint_name: str, session: Session | None = None)[source]

High-level handle for an AutoGluon-Cloud time series inference endpoint.

Wraps a SageMaker endpoint with the AutoGluon-Cloud serializer/deserializer pair, providing a clean predict() interface. Use this to attach to an existing endpoint by name. To create a new endpoint, call autogluon.cloud.TimeSeriesFoundationModel.deploy(), which returns a TimeSeriesEndpoint already pointing at the new endpoint.

Parameters:

  • endpoint_name – Name of an existing SageMaker endpoint deployed via AutoGluon-Cloud (e.g. through autogluon.cloud.TimeSeriesFoundationModel.deploy()). The endpoint must understand the AutoGluon-Cloud request payload format.

  • session – boto3.Session used to invoke and delete the endpoint. If None, the default ambient session is used.

Methods

delete_endpoint

Delete the endpoint and its backing model + endpoint config.

predict

Run real-time prediction on the deployed endpoint.

Attributes

endpoint_name