Amazon SageMaker and Amazon Forecast endpoints will be created according to the appropriate Amazon guidelines.
See AWS resources such as the SageMaker Getting Started Guide: https://aws.amazon.com/sagemaker/getting-started/.
To use API_Request in-database function, it is assumed that you already have Amazon SageMaker model or Amazon Forecast model trained and deployed in AWS cloud, and AWS assigns endpoint ID information to these models. API_Request in-database function requires this information to connect to the model for scoring. Currently, API_Request function supports only those models and endpoints that support json or csv input format.
For training, you can either use Amazon SageMaker site or SDK to train the model, or use the tdapiclient library to train the model.
- If you use Amazon SageMaker site or SDK, you need to extract endpoint ID information and use it in API_Request in-database function.
- If you use the tdapiclient library, endpoint ID information is implicitly passed to API_Request in-database function at the time of scoring.