기계 번역으로 제공되는 번역입니다. 제공된 번역과 원본 영어의 내용이 상충하는 경우에는 영어 버전이 우선합니다. # SDK for SAP ABAP를 사용한 SageMaker AI 예제 다음 코드 예제에서는 SageMaker AI와 함께 AWS SDK for SAP ABAP를 사용하여 작업을 수행하고 일반적인 시나리오를 구현하는 방법을 보여줍니다. *작업*은 대규모 프로그램에서 발췌한 코드이며 컨텍스트에 맞춰 실행해야 합니다. 작업은 개별 서비스 함수를 직접 호출하는 방법을 보여주며, 관련 시나리오의 컨텍스트에 맞는 작업을 볼 수 있습니다. *시나리오*는 동일한 서비스 내에서 또는 다른 AWS 서비스와 결합된 상태에서 여러 함수를 직접적으로 호출하여 특정 태스크를 수행하는 방법을 보여주는 코드 예제입니다. 각 예시에는 전체 소스 코드에 대한 링크가 포함되어 있으며, 여기에서 컨텍스트에 맞춰 코드를 설정하고 실행하는 방법에 대한 지침을 찾을 수 있습니다. **Topics** + [작업](#actions) + [시나리오](#scenarios) ## 작업 ### `CreateEndpoint` 다음 코드 예시는 `CreateEndpoint`의 사용 방법을 보여줍니다. **SDK for SAP ABAP API** GitHub에 더 많은 내용이 있습니다. [AWS 코드 예 리포지토리](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/sap-abap/services/sgm#code-examples)에서 전체 예를 찾고 설정 및 실행하는 방법을 배워보세요. ``` DATA lt_production_variants TYPE /aws1/cl_sgmproductionvariant=>tt_productionvariantlist. DATA lo_production_variants TYPE REF TO /aws1/cl_sgmproductionvariant. DATA oo_ep_config_result TYPE REF TO /aws1/cl_sgmcreateendptcfgout. "Create a production variant as an ABAP object." "Identifies a model that you want to host and the resources chosen to deploy for hosting it." lo_production_variants = NEW #( iv_variantname = iv_variant_name iv_modelname = iv_model_name iv_initialinstancecount = iv_initial_instance_count iv_instancetype = iv_instance_type ). INSERT lo_production_variants INTO TABLE lt_production_variants. "Create an endpoint configuration." TRY. oo_ep_config_result = lo_sgm->createendpointconfig( iv_endpointconfigname = iv_endpoint_config_name it_productionvariants = lt_production_variants ). MESSAGE 'Endpoint configuration created.' TYPE 'I'. CATCH /aws1/cx_sgmresourcelimitexcd. MESSAGE 'You have reached the limit on the number of resources.' TYPE 'E'. ENDTRY. "Create an endpoint." TRY. oo_result = lo_sgm->createendpoint( " oo_result is returned for testing purposes. " iv_endpointconfigname = iv_endpoint_config_name iv_endpointname = iv_endpoint_name ). MESSAGE 'Endpoint created.' TYPE 'I'. CATCH /aws1/cx_sgmresourcelimitexcd. MESSAGE 'You have reached the limit on the number of resources.' TYPE 'E'. ENDTRY. ``` + API 세부 정보는 *AWS SDK for SAP ABAP API 참조*의 [CreateEndpoint](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html)를 참조하세요. ### `CreateModel` 다음 코드 예시는 `CreateModel`의 사용 방법을 보여줍니다. **SDK for SAP ABAP API** GitHub에 더 많은 내용이 있습니다. [AWS 코드 예 리포지토리](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/sap-abap/services/sgm#code-examples)에서 전체 예를 찾고 설정 및 실행하는 방법을 배워보세요. ``` DATA lo_primarycontainer TYPE REF TO /aws1/cl_sgmcontainerdefn. "Create an ABAP object for the container image based on input variables." lo_primarycontainer = NEW #( iv_image = iv_container_image iv_modeldataurl = iv_model_data_url ). "Create an Amazon SageMaker model." TRY. oo_result = lo_sgm->createmodel( " oo_result is returned for testing purposes. " iv_executionrolearn = iv_execution_role_arn iv_modelname = iv_model_name io_primarycontainer = lo_primarycontainer ). MESSAGE 'Model created.' TYPE 'I'. CATCH /aws1/cx_sgmresourcelimitexcd. MESSAGE 'You have reached the limit on the number of resources.' TYPE 'E'. ENDTRY. ``` + API에 대한 세부 정보는 *AWS SDK for SAP ABAP API 참조*의 [CreateModel](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html)을 참조하세요. ### `CreateTrainingJob` 다음 코드 예시는 `CreateTrainingJob`의 사용 방법을 보여줍니다. **SDK for SAP ABAP API** GitHub에 더 많은 내용이 있습니다. [AWS 코드 예 리포지토리](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/sap-abap/services/sgm#code-examples)에서 전체 예를 찾고 설정 및 실행하는 방법을 배워보세요. ``` DATA lo_hyperparameters_w TYPE REF TO /aws1/cl_sgmhyperparameters_w. DATA lt_hyperparameters TYPE /aws1/cl_sgmhyperparameters_w=>tt_hyperparameters. DATA lt_input_data_config TYPE /aws1/cl_sgmchannel=>tt_inputdataconfig. DATA lo_trn_channel TYPE REF TO /aws1/cl_sgmchannel. DATA lo_trn_datasource TYPE REF TO /aws1/cl_sgmdatasource. DATA lo_trn_s3datasource TYPE REF TO /aws1/cl_sgms3datasource. DATA lo_val_channel TYPE REF TO /aws1/cl_sgmchannel. DATA lo_val_datasource TYPE REF TO /aws1/cl_sgmdatasource. DATA lo_val_s3datasource TYPE REF TO /aws1/cl_sgms3datasource. DATA lo_algorithm_specification TYPE REF TO /aws1/cl_sgmalgorithmspec. DATA lo_resource_config TYPE REF TO /aws1/cl_sgmresourceconfig. DATA lo_output_data_config TYPE REF TO /aws1/cl_sgmoutputdataconfig. DATA lo_stopping_condition TYPE REF TO /aws1/cl_sgmstoppingcondition. "Create ABAP internal table for hyperparameters based on input variables." "These hyperparameters are based on the Amazon SageMaker built-in algorithm, XGBoost." lo_hyperparameters_w = NEW #( iv_value = iv_hp_max_depth ). INSERT VALUE #( key = 'max_depth' value = lo_hyperparameters_w ) INTO TABLE lt_hyperparameters. lo_hyperparameters_w = NEW #( iv_value = iv_hp_eta ). INSERT VALUE #( key = 'eta' value = lo_hyperparameters_w ) INTO TABLE lt_hyperparameters. lo_hyperparameters_w = NEW #( iv_value = iv_hp_eval_metric ). INSERT VALUE #( key = 'eval_metric' value = lo_hyperparameters_w ) INTO TABLE lt_hyperparameters. lo_hyperparameters_w = NEW #( iv_value = iv_hp_scale_pos_weight ). INSERT VALUE #( key = 'scale_pos_weight' value = lo_hyperparameters_w ) INTO TABLE lt_hyperparameters. lo_hyperparameters_w = NEW #( iv_value = iv_hp_subsample ). INSERT VALUE #( key = 'subsample' value = lo_hyperparameters_w ) INTO TABLE lt_hyperparameters. lo_hyperparameters_w = NEW #( iv_value = iv_hp_objective ). INSERT VALUE #( key = 'objective' value = lo_hyperparameters_w ) INTO TABLE lt_hyperparameters. lo_hyperparameters_w = NEW #( iv_value = iv_hp_num_round ). INSERT VALUE #( key = 'num_round' value = lo_hyperparameters_w ) INTO TABLE lt_hyperparameters. "Create ABAP objects for training data sources." lo_trn_s3datasource = NEW #( iv_s3datatype = iv_trn_data_s3datatype iv_s3datadistributiontype = iv_trn_data_s3datadistribution iv_s3uri = iv_trn_data_s3uri ). lo_trn_datasource = NEW #( io_s3datasource = lo_trn_s3datasource ). lo_trn_channel = NEW #( iv_channelname = 'train' io_datasource = lo_trn_datasource iv_compressiontype = iv_trn_data_compressiontype iv_contenttype = iv_trn_data_contenttype ). INSERT lo_trn_channel INTO TABLE lt_input_data_config. "Create ABAP objects for validation data sources." lo_val_s3datasource = NEW #( iv_s3datatype = iv_val_data_s3datatype iv_s3datadistributiontype = iv_val_data_s3datadistribution iv_s3uri = iv_val_data_s3uri ). lo_val_datasource = NEW #( io_s3datasource = lo_val_s3datasource ). lo_val_channel = NEW #( iv_channelname = 'validation' io_datasource = lo_val_datasource iv_compressiontype = iv_val_data_compressiontype iv_contenttype = iv_val_data_contenttype ). INSERT lo_val_channel INTO TABLE lt_input_data_config. "Create an ABAP object for algorithm specification." lo_algorithm_specification = NEW #( iv_trainingimage = iv_training_image iv_traininginputmode = iv_training_input_mode ). "Create an ABAP object for resource configuration." lo_resource_config = NEW #( iv_instancecount = iv_instance_count iv_instancetype = iv_instance_type iv_volumesizeingb = iv_volume_sizeingb ). "Create an ABAP object for output data configuration." lo_output_data_config = NEW #( iv_s3outputpath = iv_s3_output_path ). "Create an ABAP object for stopping condition." lo_stopping_condition = NEW #( iv_maxruntimeinseconds = iv_max_runtime_in_seconds ). "Create a training job." TRY. oo_result = lo_sgm->createtrainingjob( " oo_result is returned for testing purposes. " iv_trainingjobname = iv_training_job_name iv_rolearn = iv_role_arn it_hyperparameters = lt_hyperparameters it_inputdataconfig = lt_input_data_config io_algorithmspecification = lo_algorithm_specification io_outputdataconfig = lo_output_data_config io_resourceconfig = lo_resource_config io_stoppingcondition = lo_stopping_condition ). MESSAGE 'Training job created.' TYPE 'I'. CATCH /aws1/cx_sgmresourceinuse. MESSAGE 'Resource being accessed is in use.' TYPE 'E'. CATCH /aws1/cx_sgmresourcenotfound. MESSAGE 'Resource being accessed is not found.' TYPE 'E'. CATCH /aws1/cx_sgmresourcelimitexcd. MESSAGE 'You have reached the limit on the number of resources.' TYPE 'E'. ENDTRY. ``` + API에 대한 세부 정보는 *AWS SDK for SAP ABAP API 참조*의 [CreateTrainingJob](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html)을 참조하세요. ### `CreateTransformJob` 다음 코드 예시는 `CreateTransformJob`의 사용 방법을 보여줍니다. **SDK for SAP ABAP API** GitHub에 더 많은 내용이 있습니다. [AWS 코드 예 리포지토리](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/sap-abap/services/sgm#code-examples)에서 전체 예를 찾고 설정 및 실행하는 방법을 배워보세요. ``` DATA lo_transforminput TYPE REF TO /aws1/cl_sgmtransforminput. DATA lo_transformoutput TYPE REF TO /aws1/cl_sgmtransformoutput. DATA lo_transformresources TYPE REF TO /aws1/cl_sgmtransformresources. DATA lo_datasource TYPE REF TO /aws1/cl_sgmtransformdatasrc. DATA lo_s3datasource TYPE REF TO /aws1/cl_sgmtransforms3datasrc. "Create an ABAP object for an Amazon Simple Storage Service (Amazon S3) data source." lo_s3datasource = NEW #( iv_s3uri = iv_tf_data_s3uri iv_s3datatype = iv_tf_data_s3datatype ). "Create an ABAP object for data source." lo_datasource = NEW #( io_s3datasource = lo_s3datasource ). "Create an ABAP object for transform data source." lo_transforminput = NEW #( io_datasource = lo_datasource iv_contenttype = iv_tf_data_contenttype iv_compressiontype = iv_tf_data_compressiontype ). "Create an ABAP object for resource configuration." lo_transformresources = NEW #( iv_instancecount = iv_instance_count iv_instancetype = iv_instance_type ). "Create an ABAP object for output data configuration." lo_transformoutput = NEW #( iv_s3outputpath = iv_s3_output_path ). "Create a transform job." TRY. oo_result = lo_sgm->createtransformjob( " oo_result is returned for testing purposes. " iv_modelname = iv_tf_model_name iv_transformjobname = iv_tf_job_name io_transforminput = lo_transforminput io_transformoutput = lo_transformoutput io_transformresources = lo_transformresources ). MESSAGE 'Transform job created.' TYPE 'I'. CATCH /aws1/cx_sgmresourceinuse. MESSAGE 'Resource being accessed is in use.' TYPE 'E'. CATCH /aws1/cx_sgmresourcenotfound. MESSAGE 'Resource being accessed is not found.' TYPE 'E'. CATCH /aws1/cx_sgmresourcelimitexcd. MESSAGE 'You have reached the limit on the number of resources.' TYPE 'E'. ENDTRY. ``` + API에 대한 세부 정보는 *AWS SDK for SAP ABAP API 참조*의 [CreateTransformJob](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html)을 참조하세요. ### `DeleteEndpoint` 다음 코드 예시는 `DeleteEndpoint`의 사용 방법을 보여줍니다. **SDK for SAP ABAP API** GitHub에 더 많은 내용이 있습니다. [AWS 코드 예 리포지토리](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/sap-abap/services/sgm#code-examples)에서 전체 예를 찾고 설정 및 실행하는 방법을 배워보세요. ``` "Delete an endpoint." TRY. lo_sgm->deleteendpoint( iv_endpointname = iv_endpoint_name ). MESSAGE 'Endpoint configuration deleted.' TYPE 'I'. CATCH /aws1/cx_rt_service_generic INTO DATA(lo_endpoint_exception). DATA(lv_endpoint_error) = |"{ lo_endpoint_exception->av_err_code }" - { lo_endpoint_exception->av_err_msg }|. MESSAGE lv_endpoint_error TYPE 'E'. ENDTRY. "Delete an endpoint configuration." TRY. lo_sgm->deleteendpointconfig( iv_endpointconfigname = iv_endpoint_config_name ). MESSAGE 'Endpoint deleted.' TYPE 'I'. CATCH /aws1/cx_rt_service_generic INTO DATA(lo_endpointconfig_exception). DATA(lv_endpointconfig_error) = |"{ lo_endpointconfig_exception->av_err_code }" - { lo_endpointconfig_exception->av_err_msg }|. MESSAGE lv_endpointconfig_error TYPE 'E'. ENDTRY. ``` + API 세부 정보는 *AWS SDK for SAP ABAP API 참조*의 [DeleteEndpoint](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html)를 참조하세요. ### `DeleteModel` 다음 코드 예시는 `DeleteModel`의 사용 방법을 보여줍니다. **SDK for SAP ABAP API** GitHub에 더 많은 내용이 있습니다. [AWS 코드 예 리포지토리](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/sap-abap/services/sgm#code-examples)에서 전체 예를 찾고 설정 및 실행하는 방법을 배워보세요. ``` TRY. lo_sgm->deletemodel( iv_modelname = iv_model_name ). MESSAGE 'Model deleted.' TYPE 'I'. CATCH /aws1/cx_rt_service_generic INTO DATA(lo_exception). DATA(lv_error) = |"{ lo_exception->av_err_code }" - { lo_exception->av_err_msg }|. MESSAGE lv_error TYPE 'E'. ENDTRY. ``` + API에 대한 세부 정보는 *AWS SDK for SAP ABAP API 참조*의 [DeleteModel](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html)을 참조하세요. ### `DescribeTrainingJob` 다음 코드 예시는 `DescribeTrainingJob`의 사용 방법을 보여줍니다. **SDK for SAP ABAP API** GitHub에 더 많은 내용이 있습니다. [AWS 코드 예 리포지토리](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/sap-abap/services/sgm#code-examples)에서 전체 예를 찾고 설정 및 실행하는 방법을 배워보세요. ``` TRY. oo_result = lo_sgm->describetrainingjob( " oo_result is returned for testing purposes. " iv_trainingjobname = iv_training_job_name ). MESSAGE 'Retrieved description of training job.' TYPE 'I'. CATCH /aws1/cx_rt_service_generic INTO DATA(lo_exception). DATA(lv_error) = |"{ lo_exception->av_err_code }" - { lo_exception->av_err_msg }|. MESSAGE lv_error TYPE 'E'. ENDTRY. ``` + API에 대한 세부 정보는 *AWS SDK for SAP ABAP API 참조*의 [DescribeTrainingJob](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html)을 참조하세요. ### `ListAlgorithms` 다음 코드 예시는 `ListAlgorithms`의 사용 방법을 보여줍니다. **SDK for SAP ABAP API** GitHub에 더 많은 내용이 있습니다. [AWS 코드 예 리포지토리](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/sap-abap/services/sgm#code-examples)에서 전체 예를 찾고 설정 및 실행하는 방법을 배워보세요. ``` TRY. oo_result = lo_sgm->listalgorithms( " oo_result is returned for testing purposes. " iv_namecontains = iv_name_contains ). MESSAGE 'Retrieved list of algorithms.' TYPE 'I'. CATCH /aws1/cx_rt_service_generic INTO DATA(lo_exception). DATA(lv_error) = |"{ lo_exception->av_err_code }" - { lo_exception->av_err_msg }|. MESSAGE lv_error TYPE 'E'. ENDTRY. ``` + API에 대한 세부 정보는 *AWS SDK for SAP ABAP API 참조*의 [ListAlgorithms](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html)를 참조하세요. ### `ListModels` 다음 코드 예시는 `ListModels`의 사용 방법을 보여줍니다. **SDK for SAP ABAP API** GitHub에 더 많은 내용이 있습니다. [AWS 코드 예 리포지토리](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/sap-abap/services/sgm#code-examples)에서 전체 예를 찾고 설정 및 실행하는 방법을 배워보세요. ``` TRY. oo_result = lo_sgm->listmodels( " oo_result is returned for testing purposes. " iv_namecontains = iv_name_contains ). MESSAGE 'Retrieved list of models.' TYPE 'I'. CATCH /aws1/cx_rt_service_generic INTO DATA(lo_exception). DATA(lv_error) = |"{ lo_exception->av_err_code }" - { lo_exception->av_err_msg }|. MESSAGE lv_error TYPE 'E'. ENDTRY. ``` + API에 대한 세부 정보는 *AWS SDK for SAP ABAP API 참조*의 [ListModels](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html)를 참조하세요. ### `ListNotebookInstances` 다음 코드 예시는 `ListNotebookInstances`의 사용 방법을 보여줍니다. **SDK for SAP ABAP API** GitHub에 더 많은 내용이 있습니다. [AWS 코드 예 리포지토리](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/sap-abap/services/sgm#code-examples)에서 전체 예를 찾고 설정 및 실행하는 방법을 배워보세요. ``` TRY. oo_result = lo_sgm->listnotebookinstances( " oo_result is returned for testing purposes. " iv_namecontains = iv_name_contains ). MESSAGE 'Retrieved list of notebook instances.' TYPE 'I'. CATCH /aws1/cx_rt_service_generic INTO DATA(lo_exception). DATA(lv_error) = |"{ lo_exception->av_err_code }" - { lo_exception->av_err_msg }|. MESSAGE lv_error TYPE 'E'. ENDTRY. ``` + API에 대한 세부 정보는 *AWS SDK for SAP ABAP API 참조*의 [ListNotebookInstances](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html)를 참조하세요. ### `ListTrainingJobs` 다음 코드 예시는 `ListTrainingJobs`의 사용 방법을 보여줍니다. **SDK for SAP ABAP API** GitHub에 더 많은 내용이 있습니다. [AWS 코드 예 리포지토리](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/sap-abap/services/sgm#code-examples)에서 전체 예를 찾고 설정 및 실행하는 방법을 배워보세요. ``` TRY. oo_result = lo_sgm->listtrainingjobs( " oo_result is returned for testing purposes. " iv_namecontains = iv_name_contains iv_maxresults = iv_max_results ). MESSAGE 'Retrieved list of training jobs.' TYPE 'I'. CATCH /aws1/cx_rt_service_generic INTO DATA(lo_exception). DATA(lv_error) = |"{ lo_exception->av_err_code }" - { lo_exception->av_err_msg }|. MESSAGE lv_error TYPE 'E'. ENDTRY. ``` + API에 대한 세부 정보는 *AWS SDK for SAP ABAP API 참조*의 [ListTrainingJobs](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html)를 참조하세요. ## 시나리오 ### 모델 및 엔드포인트를 사용하여 시작하기 다음 코드 예제에서는 다음과 같은 작업을 수행하는 방법을 보여줍니다. + 훈련 작업을 시작하고 SageMaker AI 모델을 만듭니다. + 엔드포인트 구성을 생성합니다. + 엔드포인트를 생성한 다음 리소스를 정리합니다. **SDK for SAP ABAP** GitHub에 더 많은 내용이 있습니다. [AWS 코드 예 리포지토리](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/sap-abap/services/sgm#code-examples)에서 전체 예를 찾고 설정 및 실행하는 방법을 배워보세요. ``` DATA lo_hyperparameters_w TYPE REF TO /aws1/cl_sgmhyperparameters_w. DATA lo_trn_channel TYPE REF TO /aws1/cl_sgmchannel. DATA lo_trn_datasource TYPE REF TO /aws1/cl_sgmdatasource. DATA lo_trn_s3datasource TYPE REF TO /aws1/cl_sgms3datasource. DATA lo_val_channel TYPE REF TO /aws1/cl_sgmchannel. DATA lo_val_datasource TYPE REF TO /aws1/cl_sgmdatasource. DATA lo_val_s3datasource TYPE REF TO /aws1/cl_sgms3datasource. DATA lo_algorithm_specification TYPE REF TO /aws1/cl_sgmalgorithmspec. DATA lo_resource_config TYPE REF TO /aws1/cl_sgmresourceconfig. DATA lo_output_data_config TYPE REF TO /aws1/cl_sgmoutputdataconfig. DATA lo_stopping_condition TYPE REF TO /aws1/cl_sgmstoppingcondition. DATA lo_primarycontainer TYPE REF TO /aws1/cl_sgmcontainerdefn. DATA lo_production_variants TYPE REF TO /aws1/cl_sgmproductionvariant. DATA lo_ep_config_result TYPE REF TO /aws1/cl_sgmcreateendptcfgout. DATA lo_training_result TYPE REF TO /aws1/cl_sgmdescrtrnjobrsp. DATA lt_production_variants TYPE /aws1/cl_sgmproductionvariant=>tt_productionvariantlist. DATA lt_input_data_config TYPE /aws1/cl_sgmchannel=>tt_inputdataconfig. DATA lt_hyperparameters TYPE /aws1/cl_sgmhyperparameters_w=>tt_hyperparameters. DATA lv_model_data_url TYPE /aws1/sgmurl. lv_model_data_url = iv_s3_output_path && iv_training_job_name && '/output/model.tar.gz'. "Create ABAP internal table for hyperparameters based on input variables." "These hyperparameters are based on Amazon SageMaker built-in algorithm - XGBoost" lo_hyperparameters_w = NEW #( iv_value = iv_hp_max_depth ). INSERT VALUE #( key = 'max_depth' value = lo_hyperparameters_w ) INTO TABLE lt_hyperparameters. lo_hyperparameters_w = NEW #( iv_value = iv_hp_eta ). INSERT VALUE #( key = 'eta' value = lo_hyperparameters_w ) INTO TABLE lt_hyperparameters. lo_hyperparameters_w = NEW #( iv_value = iv_hp_eval_metric ). INSERT VALUE #( key = 'eval_metric' value = lo_hyperparameters_w ) INTO TABLE lt_hyperparameters. lo_hyperparameters_w = NEW #( iv_value = iv_hp_scale_pos_weight ). INSERT VALUE #( key = 'scale_pos_weight' value = lo_hyperparameters_w ) INTO TABLE lt_hyperparameters. lo_hyperparameters_w = NEW #( iv_value = iv_hp_subsample ). INSERT VALUE #( key = 'subsample' value = lo_hyperparameters_w ) INTO TABLE lt_hyperparameters. lo_hyperparameters_w = NEW #( iv_value = iv_hp_objective ). INSERT VALUE #( key = 'objective' value = lo_hyperparameters_w ) INTO TABLE lt_hyperparameters. lo_hyperparameters_w = NEW #( iv_value = iv_hp_num_round ). INSERT VALUE #( key = 'num_round' value = lo_hyperparameters_w ) INTO TABLE lt_hyperparameters. "Create ABAP internal table for data based on input variables." "Training data." lo_trn_s3datasource = NEW #( iv_s3datatype = iv_trn_data_s3datatype iv_s3datadistributiontype = iv_trn_data_s3datadistribution iv_s3uri = iv_trn_data_s3uri ). lo_trn_datasource = NEW #( io_s3datasource = lo_trn_s3datasource ). lo_trn_channel = NEW #( iv_channelname = 'train' io_datasource = lo_trn_datasource iv_compressiontype = iv_trn_data_compressiontype iv_contenttype = iv_trn_data_contenttype ). INSERT lo_trn_channel INTO TABLE lt_input_data_config. "Validation data." lo_val_s3datasource = NEW #( iv_s3datatype = iv_val_data_s3datatype iv_s3datadistributiontype = iv_val_data_s3datadistribution iv_s3uri = iv_val_data_s3uri ). lo_val_datasource = NEW #( io_s3datasource = lo_val_s3datasource ). lo_val_channel = NEW #( iv_channelname = 'validation' io_datasource = lo_val_datasource iv_compressiontype = iv_val_data_compressiontype iv_contenttype = iv_val_data_contenttype ). INSERT lo_val_channel INTO TABLE lt_input_data_config. "Create an ABAP object for algorithm specification based on input variables." lo_algorithm_specification = NEW #( iv_trainingimage = iv_training_image iv_traininginputmode = iv_training_input_mode ). "Create an ABAP object for resource configuration." lo_resource_config = NEW #( iv_instancecount = iv_instance_count iv_instancetype = iv_instance_type iv_volumesizeingb = iv_volume_sizeingb ). "Create an ABAP object for output data configuration." lo_output_data_config = NEW #( iv_s3outputpath = iv_s3_output_path ). "Create an ABAP object for stopping condition." lo_stopping_condition = NEW #( iv_maxruntimeinseconds = iv_max_runtime_in_seconds ). TRY. lo_sgm->createtrainingjob( iv_trainingjobname = iv_training_job_name iv_rolearn = iv_role_arn it_hyperparameters = lt_hyperparameters it_inputdataconfig = lt_input_data_config io_algorithmspecification = lo_algorithm_specification io_outputdataconfig = lo_output_data_config io_resourceconfig = lo_resource_config io_stoppingcondition = lo_stopping_condition ). MESSAGE 'Training job created.' TYPE 'I'. CATCH /aws1/cx_sgmresourceinuse. MESSAGE 'Resource being accessed is in use.' TYPE 'E'. CATCH /aws1/cx_sgmresourcenotfound. MESSAGE 'Resource being accessed is not found.' TYPE 'E'. CATCH /aws1/cx_sgmresourcelimitexcd. MESSAGE 'You have reached the limit on the number of resources.' TYPE 'E'. ENDTRY. "Wait for training job to be completed." lo_training_result = lo_sgm->describetrainingjob( iv_trainingjobname = iv_training_job_name ). WHILE lo_training_result->get_trainingjobstatus( ) <> 'Completed'. IF sy-index = 30. EXIT. "Maximum 900 seconds." ENDIF. WAIT UP TO 30 SECONDS. lo_training_result = lo_sgm->describetrainingjob( iv_trainingjobname = iv_training_job_name ). ENDWHILE. "Create ABAP object for the container image based on input variables." lo_primarycontainer = NEW #( iv_image = iv_training_image iv_modeldataurl = lv_model_data_url ). "Create an Amazon SageMaker model." TRY. lo_sgm->createmodel( iv_executionrolearn = iv_role_arn iv_modelname = iv_model_name io_primarycontainer = lo_primarycontainer ). MESSAGE 'Model created.' TYPE 'I'. CATCH /aws1/cx_sgmresourcelimitexcd. MESSAGE 'You have reached the limit on the number of resources.' TYPE 'E'. ENDTRY. "Create an endpoint production variant." lo_production_variants = NEW #( iv_variantname = iv_ep_variant_name iv_modelname = iv_model_name iv_initialinstancecount = iv_ep_initial_instance_count iv_instancetype = iv_ep_instance_type ). INSERT lo_production_variants INTO TABLE lt_production_variants. TRY. "Create an endpoint configuration." lo_ep_config_result = lo_sgm->createendpointconfig( iv_endpointconfigname = iv_ep_cfg_name it_productionvariants = lt_production_variants ). MESSAGE 'Endpoint configuration created.' TYPE 'I'. "Create an endpoint." oo_ep_output = lo_sgm->createendpoint( " oo_ep_output is returned for testing purposes. " iv_endpointconfigname = iv_ep_cfg_name iv_endpointname = iv_ep_name ). MESSAGE 'Endpoint created.' TYPE 'I'. CATCH /aws1/cx_sgmresourcelimitexcd. MESSAGE 'You have reached the limit on the number of resources.' TYPE 'E'. ENDTRY. "Wait for endpoint creation to be completed." DATA(lo_endpoint_result) = lo_sgm->describeendpoint( iv_endpointname = iv_ep_name ). WHILE lo_endpoint_result->get_endpointstatus( ) <> 'InService'. IF sy-index = 30. EXIT. "Maximum 900 seconds." ENDIF. WAIT UP TO 30 SECONDS. lo_endpoint_result = lo_sgm->describeendpoint( iv_endpointname = iv_ep_name ). ENDWHILE. TRY. "Delete an endpoint." lo_sgm->deleteendpoint( iv_endpointname = iv_ep_name ). MESSAGE 'Endpoint deleted' TYPE 'I'. "Delete an endpoint configuration." lo_sgm->deleteendpointconfig( iv_endpointconfigname = iv_ep_cfg_name ). MESSAGE 'Endpoint configuration deleted.' TYPE 'I'. "Delete model." lo_sgm->deletemodel( iv_modelname = iv_model_name ). MESSAGE 'Model deleted.' TYPE 'I'. CATCH /aws1/cx_rt_service_generic INTO DATA(lo_endpointconfig_exception). DATA(lv_endpointconfig_error) = |"{ lo_endpointconfig_exception->av_err_code }" - { lo_endpointconfig_exception->av_err_msg }|. MESSAGE lv_endpointconfig_error TYPE 'E'. ENDTRY. ``` + API 세부 정보는 *AWS SDK for SAP ABAP API 참조*의 다음 주제를 참조하세요. + [CreateEndpoint](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html) + [CreateEndpointConfig](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html) + [CreateModel](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html) + [CreateTrainingJob](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html) + [DeleteEndpoint](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html) + [DeleteEndpointConfig](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html) + [DeleteModel](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html) + [DescribeEndpoint](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html) + [DescribeTrainingJob](https://docs.aws.amazon.com/sdk-for-sap-abap/v1/api/latest/index.html)