陰影變體 - Amazon SageMaker
部署陰影變體

本文為英文版的機器翻譯版本，如內容有任何歧義或不一致之處，概以英文版為準。

陰影變體

您可以使用 SageMaker Model Shadow 部署來建立長時間執行的陰影變體，以驗證模型服務堆疊的任何新候選元件，然後再將其升級到生產環境。下圖顯示更詳細的陰影收體運作情形。

陰影變體的詳細資訊。

部署陰影變體

下列程式碼範例示範如何以程式設計方式部署陰影變體。請將範例中的使用者預留位置文字取代為您自己的資訊。

  1. 創建兩個 SageMaker 模型：一個用於您的生產變體，另一個用於陰影變體。

    import boto3
from sagemaker import get_execution_role, Session
                
aws_region = "aws-region"

boto_session = boto3.Session(region_name=aws_region)
sagemaker_client = boto_session.client("sagemaker")

role = get_execution_role()

bucket = Session(boto_session).default_bucket()

model_name1 = "name-of-your-first-model"
model_name2 = "name-of-your-second-model"

sagemaker_client.create_model(
    ModelName = model_name1,
    ExecutionRoleArn = role,
    Containers=[
        {
            "Image": "ecr-image-uri-for-first-model",
            "ModelDataUrl": "s3-location-of-trained-first-model" 
        }
    ]
)

sagemaker_client.create_model(
    ModelName = model_name2,
    ExecutionRoleArn = role,
    Containers=[
        {
            "Image": "ecr-image-uri-for-second-model",
            "ModelDataUrl": "s3-location-of-trained-second-model" 
        }
    ]
)

  2. 建立端點組態。在組態中指定生產和陰影變體。

    endpoint_config_name = name-of-your-endpoint-config

create_endpoint_config_response = sagemaker_client.create_endpoint_config(
    EndpointConfigName=endpoint_config_name,
    ProductionVariants=[
        {
            "VariantName": name-of-your-production-variant,
            "ModelName": model_name1,
            "InstanceType": "ml.m5.xlarge",
            "InitialInstanceCount": 1,
            "InitialVariantWeight": 1,
        }
    ],
    ShadowProductionVariants=[
        {
            "VariantName": name-of-your-shadow-variant,
            "ModelName": model_name2,
            "InstanceType": "ml.m5.xlarge",
            "InitialInstanceCount": 1,
            "InitialVariantWeight": 1,
        }
   ]
)

  3. 建立端點。

    create_endpoint_response = sm.create_endpoint(
    EndpointName=name-of-your-endpoint,
    EndpointConfigName=endpoint_config_name,
)