machine-learning/triton/amazon-sagamaker-triton-resnet-50/deploy/src/undeploy.py
import boto3
def undeploy() -> None:
model_name = "resnet-50"
sagemaker_client = boto3.client(service_name="sagemaker")
sagemaker_model_name = f"{model_name}-model"
sagemaker_endpoint_config_name = f"{model_name}-endpoint-config"
sagemaker_endpoint_name = f"{model_name}-endpoint"
sagemaker_client.delete_model(ModelName=sagemaker_model_name)
sagemaker_client.delete_endpoint_config(
EndpointConfigName=sagemaker_endpoint_config_name
)
sagemaker_client.delete_endpoint(EndpointName=sagemaker_endpoint_name)
if __name__ == "__main__":
undeploy()