Model and Endpoint Management
- CreateModel
ModelName: The name of the model.
PrimaryContainer: The location of the model artifacts and Docker image for inference.
ExecutionRoleArn: The IAM role ARN that SageMaker assumes to access model artifacts and Docker images.
- DeleteModel
ModelName: The name of the model to delete.
- CreateEndpoint
EndpointName: The name of the endpoint.
EndpointConfigName: The name of the endpoint configuration to deploy.
- DeleteEndpoint
EndpointName: The name of the endpoint to delete.
- CreateEndpointConfig
EndpointConfigName: The name of the endpoint configuration.
ProductionVariants: A list of production variants (models) to deploy.
- DeleteEndpointConfig
EndpointConfigName: The name of the endpoint configuration to delete.
- UpdateEndpoint
EndpointName: The name of the existing endpoint.
EndpointConfigName: The name of the new endpoint configuration to use.
- UpdateEndpointWeightsAndCapacities
EndpointName: The name of the endpoint.
DesiredWeightsAndCapacities: A list of desired weights and capacities for each variant.
- DescribeEndpoint
EndpointName: The name of the endpoint.
- DescribeEndpointConfig
EndpointConfigName: The name of the endpoint configuration.
- ListEndpoints
MaxResults, NameContains: Optional parameters to filter the list of endpoints.
- ListEndpointConfigs
MaxResults, NameContains: Optional parameters to filter the list of endpoint configurations.
- ListModels
MaxResults, NameContains: Optional parameters to filter the list of models.
Data Preparation and Processing
- CreateProcessingJob
ProcessingJobName: The name of the processing job.
AppSpecification: The app specification that configures the processing job.
ProcessingInputs: Input data for the processing job.
ProcessingOutputConfig: Outputs configuration for the processing job.
- ListProcessingJobs
MaxResults, NameContains: Optional parameters to filter the list of processing jobs.
Training and Tuning
- CreateTrainingJob
TrainingJobName: The name of the training job.
AlgorithmSpecification: The specification of the algorithm for training.
InputDataConfig: The input data configuration for training.
OutputDataConfig: The output data configuration for training results.
- StopTrainingJob
TrainingJobName: The name of the training job to stop.