ml-pipeline-config.json

{
    "ExperimentName": "ml-pipeline-experiment",
    "ExperimentTrialsPrefix": "trial",
    "DataLakeS3Uri": "s3://{}/mlops-staging",
    "GlueArgs": {
        "--S3_SOURCE": "s3://{}/mlops-staging/sample-datasets/customer-churn.csv",
        "--S3_DEST": "s3a://{}/mlops-staging/",
        "--TRAIN_KEY": "train/",
        "--VAL_KEY": "validation/"},
    "TrainingJobName": "churn-model-xgboost",
    "ModelS3Path": "{}/model",
    "TrainDataS3Path": "{}/train/",
    "ValidationDataS3Path": "{}/validation/",
    "DataCaptureS3Path": "{}/capture",
    "MLPipelineControllerName": "ml-pipeline-controller",
    "ModelName": "customer-churn-xgb",
    "InferenceEPName": "customer-churn-xgb-ep",
    "InfInstanceType":  "ml.m5.xlarge",
    "InfBaseInstanceCount": 1,
    "TrainingImageTag": "xgboost:1",
    "TrainingInstanceType": "ml.m5.xlarge",
    "TrainingInstanceCount": 1,
    "HyperParameters": {
        "max_depth": "5",
        "eta": "0.2",
        "gamma": "4",
        "min_child_weight": "6",
        "subsample": "0.8",
        "silent": "0",
        "objective": "binary:logistic",
        "eval_metric": "error",
        "num_round": "100"
    },
    "TrainingJobParameters":{
        "AlgorithmSpecification": {
        "TrainingImage.$": "$$.Execution.Input['TrainingImage']",
        "TrainingInputMode": "File"
        },
        "OutputDataConfig": {
            "S3OutputPath.$": "$$.Execution.Input['ModelS3Path']"
        },
        "StoppingCondition": {
            "MaxRuntimeInSeconds": 86400
        },
        "ResourceConfig": {
            "InstanceCount.$": "$$.Execution.Input['TrainingInstanceCount']",
            "InstanceType.$": "$$.Execution.Input['TrainingInstanceType']",
            "VolumeSizeInGB": 30
        },
        "RoleArn.$": "$$.Execution.Input['SageMakerExecutionRole']",
        "InputDataConfig": [
            {
                "DataSource": {
                    "S3DataSource": {
                        "S3DataType": "S3Prefix",
                        "S3Uri.$": "$$.Execution.Input['TrainingDataS3Path']",
                        "S3DataDistributionType": "FullyReplicated"
                    }
                },
                "ContentType": "csv",
                "ChannelName": "train"
            },
            {
                "DataSource": {
                    "S3DataSource": {
                        "S3DataType": "S3Prefix",
                        "S3Uri.$": "$$.Execution.Input['ValidationDataS3Path']",
                        "S3DataDistributionType": "FullyReplicated"
                    }
                },
                "ContentType": "csv",
                "ChannelName": "validation"
            }
        ],
        "HyperParameters.$": "$$.Execution.Input['HyperParameters']",
        "TrainingJobName.$": "$$.Execution.Input['TrainingJobName']"
    }
}