共用方式為


Microsoft.MachineLearningServices 工作區/卷標Jobs 2020-09-01-preview

Bicep 資源定義

工作區/標籤作業資源類型可以使用目標作業進行部署:

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/labelingJobs 資源,請將下列 Bicep 新增至範本。

resource symbolicname 'Microsoft.MachineLearningServices/workspaces/labelingJobs@2020-09-01-preview' = {
  name: 'string'
  parent: resourceSymbolicName
  properties: {
    datasetConfiguration: {
      assetName: 'string'
      datasetVersion: 'string'
      enableIncrementalDatasetRefresh: bool
    }
    jobInstructions: {
      uri: 'string'
    }
    labelCategories: {
      {customized property}: {
        allowMultiSelect: bool
        classes: {
          {customized property}: {
            displayName: 'string'
            subclasses: {
              {customized property}: {}
          }
        }
        displayName: 'string'
      }
    }
    labelingJobMediaProperties: {
      annotationType: 'string'
      mediaType: 'string'
      {customized property}: any()
    }
    mlAssistConfiguration: {
      inferencingComputeBinding: {
        computeId: 'string'
        nodeCount: int
      }
      mlAssistEnabled: bool
      modelNamePrefix: 'string'
      prelabelAccuracyThreshold: int
      trainingComputeBinding: {
        computeId: 'string'
        nodeCount: int
      }
    }
    properties: {
    tags: {
      {customized property}: 'string'
    }
    {customized property}: 'string'
  }
}

屬性值

workspaces/labelingJobs

名稱 描述
NAME 資源名稱

瞭解如何在 Bicep 中設定子資源的名稱和類型。
需要字串 ()
父系 (parent) 在 Bicep 中,您可以指定子資源的父資源。 只有在父資源外部宣告子資源時,才需要新增此屬性。

如需詳細資訊,請參閱 父資源外部的子資源
類型資源的符號名稱: 工作區
properties 標籤作業的定義。 LabelingJobProperties

LabelingJobProperties

名稱 描述
datasetConfiguration 作業的數據集組態。 LabelingDatasetConfiguration (必要)
jobInstructions 作業的指示。 TagJobInstructions (必要)
labelCategories 作業的標籤類別。 LabelJobPropertiesLabelCategories (必要)
labelingJobMediaProperties 卷標作業中的媒體特定屬性。 LabelingJobImageProperties (必要)
mlAssistConfiguration 作業的機器學習輔助設定。 MLAssistConfiguration
properties job 屬性字典。 您可以新增屬性,但無法移除或改變。 LabelingJobProperties
tags 作業標籤字典。 標記可以新增、移除和更新。 物件 (object)
{自定義屬性} 字串

LabelingDatasetConfiguration

名稱 描述
assetName 要執行標籤的數據資產名稱。 需要字串 ()
datasetVersion AML 數據集版本。 需要字串 ()
enableIncrementalDatasetRefresh 指出是否要啟用累加數據集重新整理。 bool

TagJobInstructions

名稱 描述
uri 具有標籤者詳細標籤指示的頁面連結。 字串

LabelJobPropertiesLabelCategories

名稱 描述
{自定義屬性} LabelCategory

LabelCategory

名稱 描述
allowMultiSelect 指出是否允許在此類別中選取多個類別。 bool
類別 此類別中標籤類別的字典。 LabelCategoryClasses (必要)
displayName 標籤類別的顯示名稱。 字串

LabelCategoryClasses

名稱 描述
{自定義屬性} LabelClass

LabelClass

名稱 描述
displayName 標籤類別的顯示名稱。 字串
卷標類別之子類別的字典。 LabelClassSubclasses

LabelClassSubclasses

名稱 描述
{自定義屬性} LabelClass

LabelingJobImageProperties

名稱 描述
annotationType 影像標籤工作的註釋類型。 'BoundingBox'
'Classification'
'InstanceSegmentation'
mediaType 數據資產的媒體類型。 'Image'
'Text' (必要)
{自定義屬性} 針對 Bicep,您可以使用 any () 函式。

MLAssistConfiguration

名稱 描述
inferencingComputeBinding 指定用於推斷的計算。 需要 ComputeBinding ()
mlAssistEnabled 指出是否啟用 MLAssist 功能。 bool
modelNamePrefix 用於機器學習模型的名稱前置詞。 針對每個反覆專案 modelName,都會附加反覆專案,例如{modelName}_{i}。 需要字串 ()
prelabelAccuracyThreshold MLAssist 功能中使用的預先標記精確度閾值。 int
trainingComputeBinding 指定用於定型的計算。 需要 ComputeBinding ()

ComputeBinding

名稱 描述
computeId 計算資源的識別碼。 字串
nodeCount 節點數目。 int

ARM 範本資源定義

工作區/標籤作業資源類型可以使用目標作業進行部署:

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/labelingJobs 資源,請將下列 JSON 新增至範本。

{
  "type": "Microsoft.MachineLearningServices/workspaces/labelingJobs",
  "apiVersion": "2020-09-01-preview",
  "name": "string",
  "properties": {
    "datasetConfiguration": {
      "assetName": "string",
      "datasetVersion": "string",
      "enableIncrementalDatasetRefresh": "bool"
    },
    "jobInstructions": {
      "uri": "string"
    },
    "labelCategories": {
      "{customized property}": {
        "allowMultiSelect": "bool",
        "classes": {
          "{customized property}": {
            "displayName": "string",
            "subclasses": {
              "{customized property}": {}
          }
        },
        "displayName": "string"
      }
    },
    "labelingJobMediaProperties": {
      "annotationType": "string",
      "mediaType": "string",
      "{customized property}": {}
    },
    "mlAssistConfiguration": {
      "inferencingComputeBinding": {
        "computeId": "string",
        "nodeCount": "int"
      },
      "mlAssistEnabled": "bool",
      "modelNamePrefix": "string",
      "prelabelAccuracyThreshold": "int",
      "trainingComputeBinding": {
        "computeId": "string",
        "nodeCount": "int"
      }
    },
    "properties": {
    "tags": {
      "{customized property}": "string"
    },
    "{customized property}": "string"
  }
}

屬性值

workspaces/labelingJobs

名稱 描述
類型 資源類型 'Microsoft.MachineLearningServices/workspaces/labelingJobs'
apiVersion 資源 API 版本 '2020-09-01-preview'
NAME 資源名稱

瞭解如何在 JSON ARM 範本中設定子資源的名稱和類型。
需要字串 ()
properties 標籤作業的定義。 LabelingJobProperties

LabelingJobProperties

名稱 描述
datasetConfiguration 作業的數據集組態。 LabelingDatasetConfiguration (必要)
jobInstructions 作業的指示。 TagJobInstructions (必要)
labelCategories 作業的標籤類別。 LabelJobPropertiesLabelCategories (必要)
labelingJobMediaProperties 卷標作業中的媒體特定屬性。 LabelingJobImageProperties (必要)
mlAssistConfiguration 作業的機器學習輔助設定。 MLAssistConfiguration
properties job 屬性字典。 您可以新增屬性,但無法移除或改變。 LabelingJobProperties
tags 作業標籤字典。 標記可以新增、移除和更新。 物件 (object)
{自定義屬性} 字串

LabelingDatasetConfiguration

名稱 描述
assetName 要執行標籤的數據資產名稱。 需要字串 ()
datasetVersion AML 數據集版本。 需要字串 ()
enableIncrementalDatasetRefresh 指出是否要啟用累加數據集重新整理。 bool

TagJobInstructions

名稱 描述
uri 具有標籤者詳細標籤指示的頁面連結。 字串

LabelJobPropertiesLabelCategories

名稱 描述
{自定義屬性} LabelCategory

LabelCategory

名稱 描述
allowMultiSelect 指出是否允許在此類別中選取多個類別。 bool
類別 此類別中標籤類別的字典。 LabelCategoryClasses (必要)
displayName 標籤類別的顯示名稱。 字串

LabelCategoryClasses

名稱 描述
{自定義屬性} LabelClass

LabelClass

名稱 描述
displayName 標籤類別的顯示名稱。 字串
卷標類別之子類別的字典。 LabelClassSubclasses

LabelClassSubclasses

名稱 描述
{自定義屬性} LabelClass

LabelingJobImageProperties

名稱 描述
annotationType 影像標籤工作的註釋類型。 'BoundingBox'
'Classification'
'InstanceSegmentation'
mediaType 數據資產的媒體類型。 'Image'
'Text' (必要)
{自定義屬性}

MLAssistConfiguration

名稱 描述
inferencingComputeBinding 指定用於推斷的計算。 需要 ComputeBinding ()
mlAssistEnabled 指出是否啟用 MLAssist 功能。 bool
modelNamePrefix 用於機器學習模型的名稱前置詞。 針對每個反覆專案 modelName,都會附加反覆專案,例如{modelName}_{i}。 字串 (必要)
prelabelAccuracyThreshold MLAssist 功能中使用的預先標記精確度臨界值。 int
trainingComputeBinding 指定用於定型的計算。 需要 ComputeBinding ()

ComputeBinding

名稱 描述
computeId 計算資源的識別碼。 字串
nodeCount 節點數目。 int

Terraform (AzAPI 提供者) 資源定義

工作區/標籤作業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

資源格式

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

resource "azapi_resource" "symbolicname" {
  type = "Microsoft.MachineLearningServices/workspaces/labelingJobs@2020-09-01-preview"
  name = "string"
  parent_id = "string"
  body = jsonencode({
    properties = {
      datasetConfiguration = {
        assetName = "string"
        datasetVersion = "string"
        enableIncrementalDatasetRefresh = bool
      }
      jobInstructions = {
        uri = "string"
      }
      labelCategories = {
        {customized property} = {
          allowMultiSelect = bool
          classes = {
            {customized property} = {
              displayName = "string"
              subclasses = {
                {customized property} = {}
            }
          }
          displayName = "string"
        }
      }
      labelingJobMediaProperties = {
        annotationType = "string"
        mediaType = "string"
      }
      mlAssistConfiguration = {
        inferencingComputeBinding = {
          computeId = "string"
          nodeCount = int
        }
        mlAssistEnabled = bool
        modelNamePrefix = "string"
        prelabelAccuracyThreshold = int
        trainingComputeBinding = {
          computeId = "string"
          nodeCount = int
        }
      }
      properties = {
      tags = {
        {customized property} = "string"
      }
      {customized property} = "string"
    }
  })
}

屬性值

workspaces/labelingJobs

名稱 描述
類型 資源類型 “Microsoft.MachineLearningServices/workspaces/labelingJobs@2020-09-01-preview”
NAME 資源名稱 字串 (必要)
parent_id 此資源之父系的資源標識碼。 類型資源的標識碼: 工作區
properties 標籤作業的定義。 LabelingJobProperties

LabelingJobProperties

名稱 描述
datasetConfiguration 作業的數據集組態。 LabelingDatasetConfiguration (必要)
jobInstructions 作業的指示。 TagJobInstructions (必要)
labelCategories 作業的標籤類別。 LabelJobPropertiesLabelCategories (必要)
labelingJobMediaProperties 卷標作業中的媒體特定屬性。 LabelingJobImageProperties (必要)
mlAssistConfiguration 適用於作業的機器學習輔助設定。 MLAssistConfiguration
properties 作業屬性字典。 您可以新增屬性,但無法移除或改變。 LabelingJobProperties
tags 作業標籤字典。 標記可以新增、移除和更新。 物件 (object)
{自定義屬性} 字串

LabelingDatasetConfiguration

名稱 描述
assetName 要執行標籤的數據資產名稱。 字串 (必要)
datasetVersion AML 數據集版本。 字串 (必要)
enableIncrementalDatasetRefresh 指出是否要啟用累加數據集重新整理。 bool

TagJobInstructions

名稱 描述
uri 具有標籤器詳細標籤指示的頁面連結。 字串

LabelJobPropertiesLabelCategories

名稱 描述
{自定義屬性} LabelCategory

LabelCategory

名稱 描述
allowMultiSelect 指出是否允許選取此類別中的多個類別。 bool
類別 此類別中標籤類別的字典。 LabelCategoryClasses (必要)
displayName 標籤類別的顯示名稱。 字串

LabelCategoryClasses

名稱 描述
{自定義屬性} LabelClass

LabelClass

名稱 描述
displayName 標籤類別的顯示名稱。 字串
卷標類別之子類別的字典。 LabelClassSubclasses

LabelClassSubclasses

名稱 描述
{自定義屬性} LabelClass

LabelingJobImageProperties

名稱 描述
annotationType 影像標籤工作的註釋類型。 “BoundingBox”
“Classification”
“InstanceSegmentation”
mediaType 數據資產的媒體類型。 “Image”
必要) “文字” (
{自定義屬性}

MLAssistConfiguration

名稱 描述
inferencingComputeBinding 指定用於推斷的計算。 需要 ComputeBinding ()
mlAssistEnabled 指出是否啟用 MLAssist 功能。 bool
modelNamePrefix 用於機器學習模型的名稱前置詞。 針對每個反覆專案 modelName,都會附加反覆專案,例如{modelName}_{i}。 字串 (必要)
prelabelAccuracyThreshold MLAssist 功能中使用的預先標記精確度臨界值。 int
trainingComputeBinding 指定用於定型的計算。 需要 ComputeBinding ()

ComputeBinding

名稱 描述
computeId 計算資源的識別碼。 字串
nodeCount 節點數目。 int