Microsoft.MachineLearningServices 工作區/服務 2021-01-01

Bicep 資源定義

工作區/服務資源類型可以使用目標作業來部署：

• 資源群組 - 請參閱 資源群組部署命令

如需每個 API 版本中已變更屬性的清單，請參閱 變更記錄檔。

資源格式

若要建立 Microsoft.MachineLearningServices/workspaces/services 資源，請將下列 Bicep 新增至您的範本。

resource symbolicname 'Microsoft.MachineLearningServices/workspaces/services@2021-01-01' = {
  name: 'string'
  location: 'string'
  parent: resourceSymbolicName
  description: 'string'
  environmentImageRequest: {
    assets: [
      {
        id: 'string'
        mimeType: 'string'
        unpack: bool
        url: 'string'
      }
    ]
    driverProgram: 'string'
    environment: {
      docker: {
        baseDockerfile: 'string'
        baseImage: 'string'
        baseImageRegistry: {
          address: 'string'
          password: 'string'
          username: 'string'
        }
      }
      environmentVariables: {
        {customized property}: 'string'
      }
      inferencingStackVersion: 'string'
      name: 'string'
      python: {
        baseCondaEnvironment: 'string'
        condaDependencies: any()
        interpreterPath: 'string'
        userManagedDependencies: bool
      }
      r: {
        bioConductorPackages: [
          'string'
        ]
        cranPackages: [
          {
            name: 'string'
            repository: 'string'
          }
        ]
        customUrlPackages: [
          'string'
        ]
        gitHubPackages: [
          {
            authToken: 'string'
            repository: 'string'
          }
        ]
        rscriptPath: 'string'
        rVersion: 'string'
        snapshotDate: 'string'
        userManaged: bool
      }
      spark: {
        packages: [
          {
            artifact: 'string'
            group: 'string'
            version: 'string'
          }
        ]
        precachePackages: bool
        repositories: [
          'string'
        ]
      }
      version: 'string'
    }
    environmentReference: {
      name: 'string'
      version: 'string'
    }
    modelIds: [
      'string'
    ]
    models: [
      {
        createdTime: 'string'
        datasets: [
          {
            id: 'string'
            name: 'string'
          }
        ]
        derivedModelIds: [
          'string'
        ]
        description: 'string'
        experimentName: 'string'
        framework: 'string'
        frameworkVersion: 'string'
        id: 'string'
        kvTags: {}
        mimeType: 'string'
        modifiedTime: 'string'
        name: 'string'
        parentModelId: 'string'
        properties: {
          {customized property}: 'string'
        }
        resourceRequirements: {
          cpu: int
          cpuLimit: int
          fpga: int
          gpu: int
          memoryInGB: int
          memoryInGBLimit: int
        }
        runId: 'string'
        sampleInputData: 'string'
        sampleOutputData: 'string'
        unpack: bool
        url: 'string'
        version: int
      }
    ]
  }
  keys: {
    primaryKey: 'string'
    secondaryKey: 'string'
  }
  kvTags: {}
  properties: {
    {customized property}: 'string'
  }
  computeType: 'string'
  // For remaining properties, see workspaces/services objects
}

workspaces/services 物件

設定 computeType 屬性，以指定物件的類型。

針對 ACI，請使用：

  computeType: 'ACI'
  appInsightsEnabled: bool
  authEnabled: bool
  cname: 'string'
  containerResourceRequirements: {
    cpu: int
    cpuLimit: int
    fpga: int
    gpu: int
    memoryInGB: int
    memoryInGBLimit: int
  }
  dataCollection: {
    eventHubEnabled: bool
    storageEnabled: bool
  }
  dnsNameLabel: 'string'
  encryptionProperties: {
    keyName: 'string'
    keyVersion: 'string'
    vaultBaseUrl: 'string'
  }
  sslCertificate: 'string'
  sslEnabled: bool
  sslKey: 'string'
  vnetConfiguration: {
    subnetName: 'string'
    vnetName: 'string'
  }

針對 AKS，請使用：

  computeType: 'AKS'
  aadAuthEnabled: bool
  appInsightsEnabled: bool
  authEnabled: bool
  autoScaler: {
    autoscaleEnabled: bool
    maxReplicas: int
    minReplicas: int
    refreshPeriodInSeconds: int
    targetUtilization: int
  }
  computeName: 'string'
  containerResourceRequirements: {
    cpu: int
    cpuLimit: int
    fpga: int
    gpu: int
    memoryInGB: int
    memoryInGBLimit: int
  }
  dataCollection: {
    eventHubEnabled: bool
    storageEnabled: bool
  }
  isDefault: bool
  livenessProbeRequirements: {
    failureThreshold: int
    initialDelaySeconds: int
    periodSeconds: int
    successThreshold: int
    timeoutSeconds: int
  }
  maxConcurrentRequestsPerContainer: int
  maxQueueWaitMs: int
  namespace: 'string'
  numReplicas: int
  scoringTimeoutMs: int
  trafficPercentile: int
  type: 'string'

屬性值

workspaces/services

ACIServiceCreateRequest

ContainerResourceRequirements

ACIServiceCreateRequestDataCollection

ACIServiceCreateRequestEncryptionProperties

ACIServiceCreateRequestVnetConfiguration

AKSServiceCreateRequest

AKSServiceCreateRequestAutoScaler

AKSServiceCreateRequestDataCollection

AKSServiceCreateRequestLivenessProbeRequirements

CreateServiceRequestEnvironmentImageRequest

ImageAsset

EnvironmentImageRequestEnvironment

ModelEnvironmentDefinitionDocker

ModelDockerSectionBaseImageRegistry

ModelEnvironmentDefinitionEnvironmentVariables

ModelEnvironmentDefinitionPython

ModelEnvironmentDefinitionR

RCranPackage

RGitHubPackage

ModelEnvironmentDefinitionSpark

SparkMavenPackage

EnvironmentImageRequestEnvironmentReference

型

DatasetReference

ModelProperties

CreateServiceRequestKeys

CreateServiceRequestProperties

快速入門範本

下列快速入門範本會部署此資源類型。

範本	描述
建立 Azure Machine Learning Service	此範本會建立 Azure Machine Learning 服務。
建立 Azure Machine Learning 服務 （AKS）	此範本會建立 Azure Machine Learning 服務。

ARM 樣本資源定義

工作區/服務資源類型可以使用目標作業來部署：

• 資源群組 - 請參閱 資源群組部署命令

如需每個 API 版本中已變更屬性的清單，請參閱 變更記錄檔。

資源格式

若要建立 Microsoft.MachineLearningServices/workspaces/services 資源，請將下列 JSON 新增至範本。

{
  "type": "Microsoft.MachineLearningServices/workspaces/services",
  "apiVersion": "2021-01-01",
  "name": "string",
  "location": "string",
  "description": "string",
  "environmentImageRequest": {
    "assets": [
      {
        "id": "string",
        "mimeType": "string",
        "unpack": "bool",
        "url": "string"
      }
    ],
    "driverProgram": "string",
    "environment": {
      "docker": {
        "baseDockerfile": "string",
        "baseImage": "string",
        "baseImageRegistry": {
          "address": "string",
          "password": "string",
          "username": "string"
        }
      },
      "environmentVariables": {
        "{customized property}": "string"
      },
      "inferencingStackVersion": "string",
      "name": "string",
      "python": {
        "baseCondaEnvironment": "string",
        "condaDependencies": {},
        "interpreterPath": "string",
        "userManagedDependencies": "bool"
      },
      "r": {
        "bioConductorPackages": [ "string" ],
        "cranPackages": [
          {
            "name": "string",
            "repository": "string"
          }
        ],
        "customUrlPackages": [ "string" ],
        "gitHubPackages": [
          {
            "authToken": "string",
            "repository": "string"
          }
        ],
        "rscriptPath": "string",
        "rVersion": "string",
        "snapshotDate": "string",
        "userManaged": "bool"
      },
      "spark": {
        "packages": [
          {
            "artifact": "string",
            "group": "string",
            "version": "string"
          }
        ],
        "precachePackages": "bool",
        "repositories": [ "string" ]
      },
      "version": "string"
    },
    "environmentReference": {
      "name": "string",
      "version": "string"
    },
    "modelIds": [ "string" ],
    "models": [
      {
        "createdTime": "string",
        "datasets": [
          {
            "id": "string",
            "name": "string"
          }
        ],
        "derivedModelIds": [ "string" ],
        "description": "string",
        "experimentName": "string",
        "framework": "string",
        "frameworkVersion": "string",
        "id": "string",
        "kvTags": {},
        "mimeType": "string",
        "modifiedTime": "string",
        "name": "string",
        "parentModelId": "string",
        "properties": {
          "{customized property}": "string"
        },
        "resourceRequirements": {
          "cpu": "int",
          "cpuLimit": "int",
          "fpga": "int",
          "gpu": "int",
          "memoryInGB": "int",
          "memoryInGBLimit": "int"
        },
        "runId": "string",
        "sampleInputData": "string",
        "sampleOutputData": "string",
        "unpack": "bool",
        "url": "string",
        "version": "int"
      }
    ]
  },
  "keys": {
    "primaryKey": "string",
    "secondaryKey": "string"
  },
  "kvTags": {},
  "properties": {
    "{customized property}": "string"
  },
  "computeType": "string"
  // For remaining properties, see workspaces/services objects
}

workspaces/services 物件

設定 computeType 屬性，以指定物件的類型。

針對 ACI，請使用：

  "computeType": "ACI",
  "appInsightsEnabled": "bool",
  "authEnabled": "bool",
  "cname": "string",
  "containerResourceRequirements": {
    "cpu": "int",
    "cpuLimit": "int",
    "fpga": "int",
    "gpu": "int",
    "memoryInGB": "int",
    "memoryInGBLimit": "int"
  },
  "dataCollection": {
    "eventHubEnabled": "bool",
    "storageEnabled": "bool"
  },
  "dnsNameLabel": "string",
  "encryptionProperties": {
    "keyName": "string",
    "keyVersion": "string",
    "vaultBaseUrl": "string"
  },
  "sslCertificate": "string",
  "sslEnabled": "bool",
  "sslKey": "string",
  "vnetConfiguration": {
    "subnetName": "string",
    "vnetName": "string"
  }

針對 AKS，請使用：

  "computeType": "AKS",
  "aadAuthEnabled": "bool",
  "appInsightsEnabled": "bool",
  "authEnabled": "bool",
  "autoScaler": {
    "autoscaleEnabled": "bool",
    "maxReplicas": "int",
    "minReplicas": "int",
    "refreshPeriodInSeconds": "int",
    "targetUtilization": "int"
  },
  "computeName": "string",
  "containerResourceRequirements": {
    "cpu": "int",
    "cpuLimit": "int",
    "fpga": "int",
    "gpu": "int",
    "memoryInGB": "int",
    "memoryInGBLimit": "int"
  },
  "dataCollection": {
    "eventHubEnabled": "bool",
    "storageEnabled": "bool"
  },
  "isDefault": "bool",
  "livenessProbeRequirements": {
    "failureThreshold": "int",
    "initialDelaySeconds": "int",
    "periodSeconds": "int",
    "successThreshold": "int",
    "timeoutSeconds": "int"
  },
  "maxConcurrentRequestsPerContainer": "int",
  "maxQueueWaitMs": "int",
  "namespace": "string",
  "numReplicas": "int",
  "scoringTimeoutMs": "int",
  "trafficPercentile": "int",
  "type": "string"

屬性值

workspaces/services

ACIServiceCreateRequest

ContainerResourceRequirements

ACIServiceCreateRequestDataCollection

ACIServiceCreateRequestEncryptionProperties

ACIServiceCreateRequestVnetConfiguration

AKSServiceCreateRequest

AKSServiceCreateRequestAutoScaler

AKSServiceCreateRequestDataCollection

AKSServiceCreateRequestLivenessProbeRequirements

CreateServiceRequestEnvironmentImageRequest

ImageAsset

EnvironmentImageRequestEnvironment

ModelEnvironmentDefinitionDocker

ModelDockerSectionBaseImageRegistry

ModelEnvironmentDefinitionEnvironmentVariables

ModelEnvironmentDefinitionPython

ModelEnvironmentDefinitionR

RCranPackage

RGitHubPackage

ModelEnvironmentDefinitionSpark

SparkMavenPackage

EnvironmentImageRequestEnvironmentReference

型

DatasetReference

ModelProperties

CreateServiceRequestKeys

CreateServiceRequestProperties

快速入門範本

下列快速入門範本會部署此資源類型。

範本	描述
建立 Azure Machine Learning Service	此範本會建立 Azure Machine Learning 服務。
建立 Azure Machine Learning 服務 （AKS）	此範本會建立 Azure Machine Learning 服務。

Terraform （AzAPI 提供者） 資源定義

工作區/服務資源類型可以使用目標作業來部署：

• 資源群組

如需每個 API 版本中已變更屬性的清單，請參閱 變更記錄檔。

資源格式

若要建立 Microsoft.MachineLearningServices/workspaces/services 資源，請將下列 Terraform 新增至範本。

resource "azapi_resource" "symbolicname" {
  type = "Microsoft.MachineLearningServices/workspaces/services@2021-01-01"
  name = "string"
  location = "string"
  parent_id = "string"
  // For remaining properties, see workspaces/services objects
  body = jsonencode({
    description = "string"
    environmentImageRequest = {
      assets = [
        {
          id = "string"
          mimeType = "string"
          unpack = bool
          url = "string"
        }
      ]
      driverProgram = "string"
      environment = {
        docker = {
          baseDockerfile = "string"
          baseImage = "string"
          baseImageRegistry = {
            address = "string"
            password = "string"
            username = "string"
          }
        }
        environmentVariables = {
          {customized property} = "string"
        }
        inferencingStackVersion = "string"
        name = "string"
        python = {
          baseCondaEnvironment = "string"
          interpreterPath = "string"
          userManagedDependencies = bool
        }
        r = {
          bioConductorPackages = [
            "string"
          ]
          cranPackages = [
            {
              name = "string"
              repository = "string"
            }
          ]
          customUrlPackages = [
            "string"
          ]
          gitHubPackages = [
            {
              authToken = "string"
              repository = "string"
            }
          ]
          rscriptPath = "string"
          rVersion = "string"
          snapshotDate = "string"
          userManaged = bool
        }
        spark = {
          packages = [
            {
              artifact = "string"
              group = "string"
              version = "string"
            }
          ]
          precachePackages = bool
          repositories = [
            "string"
          ]
        }
        version = "string"
      }
      environmentReference = {
        name = "string"
        version = "string"
      }
      modelIds = [
        "string"
      ]
      models = [
        {
          createdTime = "string"
          datasets = [
            {
              id = "string"
              name = "string"
            }
          ]
          derivedModelIds = [
            "string"
          ]
          description = "string"
          experimentName = "string"
          framework = "string"
          frameworkVersion = "string"
          id = "string"
          kvTags = {}
          mimeType = "string"
          modifiedTime = "string"
          name = "string"
          parentModelId = "string"
          properties = {
            {customized property} = "string"
          }
          resourceRequirements = {
            cpu = int
            cpuLimit = int
            fpga = int
            gpu = int
            memoryInGB = int
            memoryInGBLimit = int
          }
          runId = "string"
          sampleInputData = "string"
          sampleOutputData = "string"
          unpack = bool
          url = "string"
          version = int
        }
      ]
    }
    keys = {
      primaryKey = "string"
      secondaryKey = "string"
    }
    kvTags = {}
    body = jsonencode({
    properties = {
      {customized property} = "string"
    }
    computeType = "string"
  })
}

workspaces/services 物件

設定 computeType 屬性，以指定物件的類型。

針對 ACI，請使用：

  computeType = "ACI"
  appInsightsEnabled = bool
  authEnabled = bool
  cname = "string"
  containerResourceRequirements = {
    cpu = int
    cpuLimit = int
    fpga = int
    gpu = int
    memoryInGB = int
    memoryInGBLimit = int
  }
  dataCollection = {
    eventHubEnabled = bool
    storageEnabled = bool
  }
  dnsNameLabel = "string"
  encryptionProperties = {
    keyName = "string"
    keyVersion = "string"
    vaultBaseUrl = "string"
  }
  sslCertificate = "string"
  sslEnabled = bool
  sslKey = "string"
  vnetConfiguration = {
    subnetName = "string"
    vnetName = "string"
  }

針對 AKS，請使用：

  computeType = "AKS"
  aadAuthEnabled = bool
  appInsightsEnabled = bool
  authEnabled = bool
  autoScaler = {
    autoscaleEnabled = bool
    maxReplicas = int
    minReplicas = int
    refreshPeriodInSeconds = int
    targetUtilization = int
  }
  computeName = "string"
  containerResourceRequirements = {
    cpu = int
    cpuLimit = int
    fpga = int
    gpu = int
    memoryInGB = int
    memoryInGBLimit = int
  }
  dataCollection = {
    eventHubEnabled = bool
    storageEnabled = bool
  }
  isDefault = bool
  livenessProbeRequirements = {
    failureThreshold = int
    initialDelaySeconds = int
    periodSeconds = int
    successThreshold = int
    timeoutSeconds = int
  }
  maxConcurrentRequestsPerContainer = int
  maxQueueWaitMs = int
  namespace = "string"
  numReplicas = int
  scoringTimeoutMs = int
  trafficPercentile = int
  type = "string"

屬性值

workspaces/services

ACIServiceCreateRequest

ContainerResourceRequirements

ACIServiceCreateRequestDataCollection

ACIServiceCreateRequestEncryptionProperties

ACIServiceCreateRequestVnetConfiguration

AKSServiceCreateRequest

AKSServiceCreateRequestAutoScaler

AKSServiceCreateRequestDataCollection

AKSServiceCreateRequestLivenessProbeRequirements

CreateServiceRequestEnvironmentImageRequest

ImageAsset

EnvironmentImageRequestEnvironment

ModelEnvironmentDefinitionDocker

ModelDockerSectionBaseImageRegistry

ModelEnvironmentDefinitionEnvironmentVariables

ModelEnvironmentDefinitionPython

ModelEnvironmentDefinitionR

RCranPackage

RGitHubPackage

ModelEnvironmentDefinitionSpark

SparkMavenPackage

EnvironmentImageRequestEnvironmentReference

型

DatasetReference

ModelProperties

CreateServiceRequestKeys

CreateServiceRequestProperties