バースト可能なインスタンスを使用したモデル・デプロイメントの作成

モデル・デプロイメントを作成するには、oci data-science model-deployment updateコマンドと必要なパラメータを使用します:

oci data-science model-deployment create --required-param-name variable-name ... [OPTIONS]

oci data-science model-deployment create \
--compartment-id <MODEL_DEPLOYMENT_COMPARTMENT_OCID> \
--model-deployment-configuration-details file://<MODEL_DEPLOYMENT_CONFIGURATION_FILE> \
--project-id <PROJECT_OCID> \
--display-name <MODEL_DEPLOYMENT_NAME>

{
  "modelDeploymentConfigurationDetails": {
    "deploymentType": "SINGLE_MODEL",
    "modelConfigurationDetails": {
      "modelId": "ocid1.datasciencemodel.oc1.iad........",
      "instanceConfiguration": {
        "instanceShapeName": "VM.Standard.E4.Flex",
        "modelDeploymentInstanceShapeConfigDetails": {
          "ocpus": 1,
          "memoryInGBs": 16,
          "cpuBaseline": "BASELINE_1_2"
        }
      },
      "scalingPolicy": {
        "policyType": "FIXED_SIZE",
        "instanceCount": 1
      },
      "bandwidthMbps": 10
    },
    "environmentConfigurationDetails" : {
      "environmentConfigurationType": "OCIR_CONTAINER",
      "image": "iad.ocir.io/testtenancy/image_name:1.0.0",
      "entrypoint": [
        "python",
        "/opt/entrypoint.py"
      ],
      "serverPort": "5000",
      "healthCheckPort": "5000"
    },
    "streamConfigurationDetails": {
      "inputStreamIds": null,
      "outputStreamIds": null
    }
  }
}

CLIコマンドのパラメータおよび値の完全なリストは、CLIコマンド・リファレンスを参照してください。

CreateModelDeployment操作を使用して、モデル・デプロイメントを作成します。

endpoint = (
    "https://datascience.eu-milan-1.oci.oraclecloud.com/20190101/modelDeployments/"
)
 
config = oci.config.from_file(profile_name="OC1")
config["region"] = "us-ashburn-1"
token_file = config.get("security_token_file", None)
signer = None
if token_file:
    with open(token_file, "r", encoding="UTF-8") as file:
        token = file.read()
    private_key = oci.signer.load_private_key_from_file(config.get("key_file", ""))
    signer = oci.auth.signers.SecurityTokenSigner(token, private_key)
 
model_deployment_details_payload = {
    "projectId": "ocid1.datascienceproject.oc1.eu-milan-1.amaaaaaabbbcccccc",
    "modelDeploymentConfigurationDetails": {
        "deploymentType": "SINGLE_MODEL",
        "modelConfigurationDetails": {
            "modelId": "ocid1.datasciencemodel.oc1.iad.amaaaaaav66vvniatytp7xjhid3inftgo52j55ozlm7hzglvvqni5wxi3bja",
            "scalingPolicy": {
                "instanceCount": 1,
                "policyType": "FIXED_SIZE"
            },
        "instanceConfiguration": {
            "instanceShapeName": "VM.Standard.E4.Flex",
            "modelDeploymentInstanceShapeConfigDetails": {
                "ocpus": 1,
                "memoryInGBs": 8,
                "cpuBaseline": "BASELINE_1_2"
        }
      }
    },
    "streamConfigurationDetails": null
  },
        "streamConfigurationDetails": None,
        "environmentConfigurationDetails": None,
    },
    "deploymentMode": "HTTPS_ONLY",
    "displayName": "MD With Burstable VM",
    "compartmentId": "ocid1.compartment.oc1..aaaaaaaabcccccccc",
}
 
create_response = requests.post(
    endpoint, auth=signer, json=model_deployment_details_payload
)
print(create_response.text)