Microsoft.MachineLearningServices workspaces/jobs 2024-04-01

Article
12/09/2024

Bicep resource definition

The workspaces/jobs resource type can be deployed with operations that target:

Resource groups - See resource group deployment commands

For a list of changed properties in each API version, see change log.

Resource format

To create a Microsoft.MachineLearningServices/workspaces/jobs resource, add the following Bicep to your template.

resource symbolicname 'Microsoft.MachineLearningServices/workspaces/jobs@2024-04-01' = {
  parent: resourceSymbolicName
  name: 'string'
  properties: {
    componentId: 'string'
    computeId: 'string'
    description: 'string'
    displayName: 'string'
    experimentName: 'string'
    identity: {
      identityType: 'string'
      // For remaining properties, see IdentityConfiguration objects
    }
    isArchived: bool
    notificationSetting: {
      emailOn: [
        'string'
      ]
      emails: [
        'string'
      ]
      webhooks: {
        {customized property}: {
          eventType: 'string'
          webhookType: 'string'
          // For remaining properties, see Webhook objects
        }
      }
    }
    properties: {
      {customized property}: 'string'
    }
    services: {
      {customized property}: {
        endpoint: 'string'
        jobServiceType: 'string'
        nodes: {
          nodesValueType: 'string'
          // For remaining properties, see Nodes objects
        }
        port: int
        properties: {
          {customized property}: 'string'
        }
      }
    }
    tags: {
      {customized property}: 'string'
    }
    jobType: 'string'
    // For remaining properties, see JobBaseProperties objects
  }
}

EarlyTerminationPolicy objects

Set the policyType property to specify the type of object.

For Bandit, use:

{
  policyType: 'Bandit'
  slackAmount: int
  slackFactor: int
}

For MedianStopping, use:

{
  policyType: 'MedianStopping'
}

For TruncationSelection, use:

{
  policyType: 'TruncationSelection'
  truncationPercentage: int
}

JobOutput objects

Set the jobOutputType property to specify the type of object.

For custom_model, use:

{
  jobOutputType: 'custom_model'
  mode: 'string'
  uri: 'string'
}

For mlflow_model, use:

{
  jobOutputType: 'mlflow_model'
  mode: 'string'
  uri: 'string'
}

For mltable, use:

{
  jobOutputType: 'mltable'
  mode: 'string'
  uri: 'string'
}

For triton_model, use:

{
  jobOutputType: 'triton_model'
  mode: 'string'
  uri: 'string'
}

For uri_file, use:

{
  jobOutputType: 'uri_file'
  mode: 'string'
  uri: 'string'
}

For uri_folder, use:

{
  jobOutputType: 'uri_folder'
  mode: 'string'
  uri: 'string'
}

SamplingAlgorithm objects

Set the samplingAlgorithmType property to specify the type of object.

For Bayesian, use:

{
  samplingAlgorithmType: 'Bayesian'
}

For Grid, use:

{
  samplingAlgorithmType: 'Grid'
}

For Random, use:

{
  rule: 'string'
  samplingAlgorithmType: 'Random'
  seed: int
}

Webhook objects

Set the webhookType property to specify the type of object.

For AzureDevOps, use:

{
  webhookType: 'AzureDevOps'
}

NCrossValidations objects

Set the mode property to specify the type of object.

For Auto, use:

{
  mode: 'Auto'
}

For Custom, use:

{
  mode: 'Custom'
  value: int
}

DistributionConfiguration objects

Set the distributionType property to specify the type of object.

For Mpi, use:

{
  distributionType: 'Mpi'
  processCountPerInstance: int
}

For PyTorch, use:

{
  distributionType: 'PyTorch'
  processCountPerInstance: int
}

For TensorFlow, use:

{
  distributionType: 'TensorFlow'
  parameterServerCount: int
  workerCount: int
}

TargetLags objects

Set the mode property to specify the type of object.

For Auto, use:

{
  mode: 'Auto'
}

For Custom, use:

{
  mode: 'Custom'
  values: [
    int
  ]
}

ForecastHorizon objects

Set the mode property to specify the type of object.

For Auto, use:

{
  mode: 'Auto'
}

For Custom, use:

{
  mode: 'Custom'
  value: int
}

IdentityConfiguration objects

Set the identityType property to specify the type of object.

For AMLToken, use:

{
  identityType: 'AMLToken'
}

For Managed, use:

{
  clientId: 'string'
  identityType: 'Managed'
  objectId: 'string'
  resourceId: 'string'
}

For UserIdentity, use:

{
  identityType: 'UserIdentity'
}

Seasonality objects

Set the mode property to specify the type of object.

For Auto, use:

{
  mode: 'Auto'
}

For Custom, use:

{
  mode: 'Custom'
  value: int
}

JobBaseProperties objects

Set the jobType property to specify the type of object.

For AutoML, use:

{
  environmentId: 'string'
  environmentVariables: {
    {customized property}: 'string'
  }
  jobType: 'AutoML'
  outputs: {
    {customized property}: {
      description: 'string'
      jobOutputType: 'string'
      // For remaining properties, see JobOutput objects
    }
  }
  queueSettings: {
    jobTier: 'string'
  }
  resources: {
    dockerArgs: 'string'
    instanceCount: int
    instanceType: 'string'
    properties: {
      {customized property}: any(Azure.Bicep.Types.Concrete.AnyType)
    }
    shmSize: 'string'
  }
  taskDetails: {
    logVerbosity: 'string'
    targetColumnName: 'string'
    trainingData: {
      description: 'string'
      jobInputType: 'string'
      mode: 'string'
      uri: 'string'
    }
    taskType: 'string'
    // For remaining properties, see AutoMLVertical objects
  }
}

For Command, use:

{
  codeId: 'string'
  command: 'string'
  distribution: {
    distributionType: 'string'
    // For remaining properties, see DistributionConfiguration objects
  }
  environmentId: 'string'
  environmentVariables: {
    {customized property}: 'string'
  }
  inputs: {
    {customized property}: {
      description: 'string'
      jobInputType: 'string'
      // For remaining properties, see JobInput objects
    }
  }
  jobType: 'Command'
  limits: {
    jobLimitsType: 'string'
    timeout: 'string'
  }
  outputs: {
    {customized property}: {
      description: 'string'
      jobOutputType: 'string'
      // For remaining properties, see JobOutput objects
    }
  }
  queueSettings: {
    jobTier: 'string'
  }
  resources: {
    dockerArgs: 'string'
    instanceCount: int
    instanceType: 'string'
    properties: {
      {customized property}: any(Azure.Bicep.Types.Concrete.AnyType)
    }
    shmSize: 'string'
  }
}

For Pipeline, use:

{
  inputs: {
    {customized property}: {
      description: 'string'
      jobInputType: 'string'
      // For remaining properties, see JobInput objects
    }
  }
  jobs: {
    {customized property}: any(Azure.Bicep.Types.Concrete.AnyType)
  }
  jobType: 'Pipeline'
  outputs: {
    {customized property}: {
      description: 'string'
      jobOutputType: 'string'
      // For remaining properties, see JobOutput objects
    }
  }
  settings: any(Azure.Bicep.Types.Concrete.AnyType)
  sourceJobId: 'string'
}

For Spark, use:

{
  archives: [
    'string'
  ]
  args: 'string'
  codeId: 'string'
  conf: {
    {customized property}: 'string'
  }
  entry: {
    sparkJobEntryType: 'string'
    // For remaining properties, see SparkJobEntry objects
  }
  environmentId: 'string'
  environmentVariables: {
    {customized property}: 'string'
  }
  files: [
    'string'
  ]
  inputs: {
    {customized property}: {
      description: 'string'
      jobInputType: 'string'
      // For remaining properties, see JobInput objects
    }
  }
  jars: [
    'string'
  ]
  jobType: 'Spark'
  outputs: {
    {customized property}: {
      description: 'string'
      jobOutputType: 'string'
      // For remaining properties, see JobOutput objects
    }
  }
  pyFiles: [
    'string'
  ]
  queueSettings: {
    jobTier: 'string'
  }
  resources: {
    instanceType: 'string'
    runtimeVersion: 'string'
  }
}

For Sweep, use:

{
  earlyTermination: {
    delayEvaluation: int
    evaluationInterval: int
    policyType: 'string'
    // For remaining properties, see EarlyTerminationPolicy objects
  }
  inputs: {
    {customized property}: {
      description: 'string'
      jobInputType: 'string'
      // For remaining properties, see JobInput objects
    }
  }
  jobType: 'Sweep'
  limits: {
    jobLimitsType: 'string'
    maxConcurrentTrials: int
    maxTotalTrials: int
    timeout: 'string'
    trialTimeout: 'string'
  }
  objective: {
    goal: 'string'
    primaryMetric: 'string'
  }
  outputs: {
    {customized property}: {
      description: 'string'
      jobOutputType: 'string'
      // For remaining properties, see JobOutput objects
    }
  }
  queueSettings: {
    jobTier: 'string'
  }
  samplingAlgorithm: {
    samplingAlgorithmType: 'string'
    // For remaining properties, see SamplingAlgorithm objects
  }
  searchSpace: any(Azure.Bicep.Types.Concrete.AnyType)
  trial: {
    codeId: 'string'
    command: 'string'
    distribution: {
      distributionType: 'string'
      // For remaining properties, see DistributionConfiguration objects
    }
    environmentId: 'string'
    environmentVariables: {
      {customized property}: 'string'
    }
    resources: {
      dockerArgs: 'string'
      instanceCount: int
      instanceType: 'string'
      properties: {
        {customized property}: any(Azure.Bicep.Types.Concrete.AnyType)
      }
      shmSize: 'string'
    }
  }
}

Nodes objects

Set the nodesValueType property to specify the type of object.

For All, use:

{
  nodesValueType: 'All'
}

SparkJobEntry objects

Set the sparkJobEntryType property to specify the type of object.

For SparkJobPythonEntry, use:

{
  file: 'string'
  sparkJobEntryType: 'SparkJobPythonEntry'
}

For SparkJobScalaEntry, use:

{
  className: 'string'
  sparkJobEntryType: 'SparkJobScalaEntry'
}

TargetRollingWindowSize objects

Set the mode property to specify the type of object.

For Auto, use:

{
  mode: 'Auto'
}

For Custom, use:

{
  mode: 'Custom'
  value: int
}

AutoMLVertical objects

Set the taskType property to specify the type of object.

For Classification, use:

{
  cvSplitColumnNames: [
    'string'
  ]
  featurizationSettings: {
    blockedTransformers: [
      'string'
    ]
    columnNameAndTypes: {
      {customized property}: 'string'
    }
    datasetLanguage: 'string'
    enableDnnFeaturization: bool
    mode: 'string'
    transformerParams: {
      {customized property}: [
        {
          fields: [
            'string'
          ]
          parameters: any(Azure.Bicep.Types.Concrete.AnyType)
        }
      ]
    }
  }
  limitSettings: {
    enableEarlyTermination: bool
    exitScore: int
    maxConcurrentTrials: int
    maxCoresPerTrial: int
    maxTrials: int
    timeout: 'string'
    trialTimeout: 'string'
  }
  nCrossValidations: {
    mode: 'string'
    // For remaining properties, see NCrossValidations objects
  }
  positiveLabel: 'string'
  primaryMetric: 'string'
  taskType: 'Classification'
  testData: {
    description: 'string'
    jobInputType: 'string'
    mode: 'string'
    uri: 'string'
  }
  testDataSize: int
  trainingSettings: {
    allowedTrainingAlgorithms: [
      'string'
    ]
    blockedTrainingAlgorithms: [
      'string'
    ]
    enableDnnTraining: bool
    enableModelExplainability: bool
    enableOnnxCompatibleModels: bool
    enableStackEnsemble: bool
    enableVoteEnsemble: bool
    ensembleModelDownloadTimeout: 'string'
    stackEnsembleSettings: {
      stackMetaLearnerKWargs: any(Azure.Bicep.Types.Concrete.AnyType)
      stackMetaLearnerTrainPercentage: int
      stackMetaLearnerType: 'string'
    }
  }
  validationData: {
    description: 'string'
    jobInputType: 'string'
    mode: 'string'
    uri: 'string'
  }
  validationDataSize: int
  weightColumnName: 'string'
}

For Forecasting, use:

{
  cvSplitColumnNames: [
    'string'
  ]
  featurizationSettings: {
    blockedTransformers: [
      'string'
    ]
    columnNameAndTypes: {
      {customized property}: 'string'
    }
    datasetLanguage: 'string'
    enableDnnFeaturization: bool
    mode: 'string'
    transformerParams: {
      {customized property}: [
        {
          fields: [
            'string'
          ]
          parameters: any(Azure.Bicep.Types.Concrete.AnyType)
        }
      ]
    }
  }
  forecastingSettings: {
    countryOrRegionForHolidays: 'string'
    cvStepSize: int
    featureLags: 'string'
    forecastHorizon: {
      mode: 'string'
      // For remaining properties, see ForecastHorizon objects
    }
    frequency: 'string'
    seasonality: {
      mode: 'string'
      // For remaining properties, see Seasonality objects
    }
    shortSeriesHandlingConfig: 'string'
    targetAggregateFunction: 'string'
    targetLags: {
      mode: 'string'
      // For remaining properties, see TargetLags objects
    }
    targetRollingWindowSize: {
      mode: 'string'
      // For remaining properties, see TargetRollingWindowSize objects
    }
    timeColumnName: 'string'
    timeSeriesIdColumnNames: [
      'string'
    ]
    useStl: 'string'
  }
  limitSettings: {
    enableEarlyTermination: bool
    exitScore: int
    maxConcurrentTrials: int
    maxCoresPerTrial: int
    maxTrials: int
    timeout: 'string'
    trialTimeout: 'string'
  }
  nCrossValidations: {
    mode: 'string'
    // For remaining properties, see NCrossValidations objects
  }
  primaryMetric: 'string'
  taskType: 'Forecasting'
  testData: {
    description: 'string'
    jobInputType: 'string'
    mode: 'string'
    uri: 'string'
  }
  testDataSize: int
  trainingSettings: {
    allowedTrainingAlgorithms: [
      'string'
    ]
    blockedTrainingAlgorithms: [
      'string'
    ]
    enableDnnTraining: bool
    enableModelExplainability: bool
    enableOnnxCompatibleModels: bool
    enableStackEnsemble: bool
    enableVoteEnsemble: bool
    ensembleModelDownloadTimeout: 'string'
    stackEnsembleSettings: {
      stackMetaLearnerKWargs: any(Azure.Bicep.Types.Concrete.AnyType)
      stackMetaLearnerTrainPercentage: int
      stackMetaLearnerType: 'string'
    }
  }
  validationData: {
    description: 'string'
    jobInputType: 'string'
    mode: 'string'
    uri: 'string'
  }
  validationDataSize: int
  weightColumnName: 'string'
}

For ImageClassification, use:

{
  limitSettings: {
    maxConcurrentTrials: int
    maxTrials: int
    timeout: 'string'
  }
  modelSettings: {
    advancedSettings: 'string'
    amsGradient: bool
    augmentations: 'string'
    beta1: int
    beta2: int
    checkpointFrequency: int
    checkpointModel: {
      description: 'string'
      jobInputType: 'string'
      mode: 'string'
      uri: 'string'
    }
    checkpointRunId: 'string'
    distributed: bool
    earlyStopping: bool
    earlyStoppingDelay: int
    earlyStoppingPatience: int
    enableOnnxNormalization: bool
    evaluationFrequency: int
    gradientAccumulationStep: int
    layersToFreeze: int
    learningRate: int
    learningRateScheduler: 'string'
    modelName: 'string'
    momentum: int
    nesterov: bool
    numberOfEpochs: int
    numberOfWorkers: int
    optimizer: 'string'
    randomSeed: int
    stepLRGamma: int
    stepLRStepSize: int
    trainingBatchSize: int
    trainingCropSize: int
    validationBatchSize: int
    validationCropSize: int
    validationResizeSize: int
    warmupCosineLRCycles: int
    warmupCosineLRWarmupEpochs: int
    weightDecay: int
    weightedLoss: int
  }
  primaryMetric: 'string'
  searchSpace: [
    {
      amsGradient: 'string'
      augmentations: 'string'
      beta1: 'string'
      beta2: 'string'
      distributed: 'string'
      earlyStopping: 'string'
      earlyStoppingDelay: 'string'
      earlyStoppingPatience: 'string'
      enableOnnxNormalization: 'string'
      evaluationFrequency: 'string'
      gradientAccumulationStep: 'string'
      layersToFreeze: 'string'
      learningRate: 'string'
      learningRateScheduler: 'string'
      modelName: 'string'
      momentum: 'string'
      nesterov: 'string'
      numberOfEpochs: 'string'
      numberOfWorkers: 'string'
      optimizer: 'string'
      randomSeed: 'string'
      stepLRGamma: 'string'
      stepLRStepSize: 'string'
      trainingBatchSize: 'string'
      trainingCropSize: 'string'
      validationBatchSize: 'string'
      validationCropSize: 'string'
      validationResizeSize: 'string'
      warmupCosineLRCycles: 'string'
      warmupCosineLRWarmupEpochs: 'string'
      weightDecay: 'string'
      weightedLoss: 'string'
    }
  ]
  sweepSettings: {
    earlyTermination: {
      delayEvaluation: int
      evaluationInterval: int
      policyType: 'string'
      // For remaining properties, see EarlyTerminationPolicy objects
    }
    samplingAlgorithm: 'string'
  }
  taskType: 'ImageClassification'
  validationData: {
    description: 'string'
    jobInputType: 'string'
    mode: 'string'
    uri: 'string'
  }
  validationDataSize: int
}

For ImageClassificationMultilabel, use:

{
  limitSettings: {
    maxConcurrentTrials: int
    maxTrials: int
    timeout: 'string'
  }
  modelSettings: {
    advancedSettings: 'string'
    amsGradient: bool
    augmentations: 'string'
    beta1: int
    beta2: int
    checkpointFrequency: int
    checkpointModel: {
      description: 'string'
      jobInputType: 'string'
      mode: 'string'
      uri: 'string'
    }
    checkpointRunId: 'string'
    distributed: bool
    earlyStopping: bool
    earlyStoppingDelay: int
    earlyStoppingPatience: int
    enableOnnxNormalization: bool
    evaluationFrequency: int
    gradientAccumulationStep: int
    layersToFreeze: int
    learningRate: int
    learningRateScheduler: 'string'
    modelName: 'string'
    momentum: int
    nesterov: bool
    numberOfEpochs: int
    numberOfWorkers: int
    optimizer: 'string'
    randomSeed: int
    stepLRGamma: int
    stepLRStepSize: int
    trainingBatchSize: int
    trainingCropSize: int
    validationBatchSize: int
    validationCropSize: int
    validationResizeSize: int
    warmupCosineLRCycles: int
    warmupCosineLRWarmupEpochs: int
    weightDecay: int
    weightedLoss: int
  }
  primaryMetric: 'string'
  searchSpace: [
    {
      amsGradient: 'string'
      augmentations: 'string'
      beta1: 'string'
      beta2: 'string'
      distributed: 'string'
      earlyStopping: 'string'
      earlyStoppingDelay: 'string'
      earlyStoppingPatience: 'string'
      enableOnnxNormalization: 'string'
      evaluationFrequency: 'string'
      gradientAccumulationStep: 'string'
      layersToFreeze: 'string'
      learningRate: 'string'
      learningRateScheduler: 'string'
      modelName: 'string'
      momentum: 'string'
      nesterov: 'string'
      numberOfEpochs: 'string'
      numberOfWorkers: 'string'
      optimizer: 'string'
      randomSeed: 'string'
      stepLRGamma: 'string'
      stepLRStepSize: 'string'
      trainingBatchSize: 'string'
      trainingCropSize: 'string'
      validationBatchSize: 'string'
      validationCropSize: 'string'
      validationResizeSize: 'string'
      warmupCosineLRCycles: 'string'
      warmupCosineLRWarmupEpochs: 'string'
      weightDecay: 'string'
      weightedLoss: 'string'
    }
  ]
  sweepSettings: {
    earlyTermination: {
      delayEvaluation: int
      evaluationInterval: int
      policyType: 'string'
      // For remaining properties, see EarlyTerminationPolicy objects
    }
    samplingAlgorithm: 'string'
  }
  taskType: 'ImageClassificationMultilabel'
  validationData: {
    description: 'string'
    jobInputType: 'string'
    mode: 'string'
    uri: 'string'
  }
  validationDataSize: int
}

For ImageInstanceSegmentation, use:

{
  limitSettings: {
    maxConcurrentTrials: int
    maxTrials: int
    timeout: 'string'
  }
  modelSettings: {
    advancedSettings: 'string'
    amsGradient: bool
    augmentations: 'string'
    beta1: int
    beta2: int
    boxDetectionsPerImage: int
    boxScoreThreshold: int
    checkpointFrequency: int
    checkpointModel: {
      description: 'string'
      jobInputType: 'string'
      mode: 'string'
      uri: 'string'
    }
    checkpointRunId: 'string'
    distributed: bool
    earlyStopping: bool
    earlyStoppingDelay: int
    earlyStoppingPatience: int
    enableOnnxNormalization: bool
    evaluationFrequency: int
    gradientAccumulationStep: int
    imageSize: int
    layersToFreeze: int
    learningRate: int
    learningRateScheduler: 'string'
    maxSize: int
    minSize: int
    modelName: 'string'
    modelSize: 'string'
    momentum: int
    multiScale: bool
    nesterov: bool
    nmsIouThreshold: int
    numberOfEpochs: int
    numberOfWorkers: int
    optimizer: 'string'
    randomSeed: int
    stepLRGamma: int
    stepLRStepSize: int
    tileGridSize: 'string'
    tileOverlapRatio: int
    tilePredictionsNmsThreshold: int
    trainingBatchSize: int
    validationBatchSize: int
    validationIouThreshold: int
    validationMetricType: 'string'
    warmupCosineLRCycles: int
    warmupCosineLRWarmupEpochs: int
    weightDecay: int
  }
  primaryMetric: 'string'
  searchSpace: [
    {
      amsGradient: 'string'
      augmentations: 'string'
      beta1: 'string'
      beta2: 'string'
      boxDetectionsPerImage: 'string'
      boxScoreThreshold: 'string'
      distributed: 'string'
      earlyStopping: 'string'
      earlyStoppingDelay: 'string'
      earlyStoppingPatience: 'string'
      enableOnnxNormalization: 'string'
      evaluationFrequency: 'string'
      gradientAccumulationStep: 'string'
      imageSize: 'string'
      layersToFreeze: 'string'
      learningRate: 'string'
      learningRateScheduler: 'string'
      maxSize: 'string'
      minSize: 'string'
      modelName: 'string'
      modelSize: 'string'
      momentum: 'string'
      multiScale: 'string'
      nesterov: 'string'
      nmsIouThreshold: 'string'
      numberOfEpochs: 'string'
      numberOfWorkers: 'string'
      optimizer: 'string'
      randomSeed: 'string'
      stepLRGamma: 'string'
      stepLRStepSize: 'string'
      tileGridSize: 'string'
      tileOverlapRatio: 'string'
      tilePredictionsNmsThreshold: 'string'
      trainingBatchSize: 'string'
      validationBatchSize: 'string'
      validationIouThreshold: 'string'
      validationMetricType: 'string'
      warmupCosineLRCycles: 'string'
      warmupCosineLRWarmupEpochs: 'string'
      weightDecay: 'string'
    }
  ]
  sweepSettings: {
    earlyTermination: {
      delayEvaluation: int
      evaluationInterval: int
      policyType: 'string'
      // For remaining properties, see EarlyTerminationPolicy objects
    }
    samplingAlgorithm: 'string'
  }
  taskType: 'ImageInstanceSegmentation'
  validationData: {
    description: 'string'
    jobInputType: 'string'
    mode: 'string'
    uri: 'string'
  }
  validationDataSize: int
}

For ImageObjectDetection, use:

{
  limitSettings: {
    maxConcurrentTrials: int
    maxTrials: int
    timeout: 'string'
  }
  modelSettings: {
    advancedSettings: 'string'
    amsGradient: bool
    augmentations: 'string'
    beta1: int
    beta2: int
    boxDetectionsPerImage: int
    boxScoreThreshold: int
    checkpointFrequency: int
    checkpointModel: {
      description: 'string'
      jobInputType: 'string'
      mode: 'string'
      uri: 'string'
    }
    checkpointRunId: 'string'
    distributed: bool
    earlyStopping: bool
    earlyStoppingDelay: int
    earlyStoppingPatience: int
    enableOnnxNormalization: bool
    evaluationFrequency: int
    gradientAccumulationStep: int
    imageSize: int
    layersToFreeze: int
    learningRate: int
    learningRateScheduler: 'string'
    maxSize: int
    minSize: int
    modelName: 'string'
    modelSize: 'string'
    momentum: int
    multiScale: bool
    nesterov: bool
    nmsIouThreshold: int
    numberOfEpochs: int
    numberOfWorkers: int
    optimizer: 'string'
    randomSeed: int
    stepLRGamma: int
    stepLRStepSize: int
    tileGridSize: 'string'
    tileOverlapRatio: int
    tilePredictionsNmsThreshold: int
    trainingBatchSize: int
    validationBatchSize: int
    validationIouThreshold: int
    validationMetricType: 'string'
    warmupCosineLRCycles: int
    warmupCosineLRWarmupEpochs: int
    weightDecay: int
  }
  primaryMetric: 'string'
  searchSpace: [
    {
      amsGradient: 'string'
      augmentations: 'string'
      beta1: 'string'
      beta2: 'string'
      boxDetectionsPerImage: 'string'
      boxScoreThreshold: 'string'
      distributed: 'string'
      earlyStopping: 'string'
      earlyStoppingDelay: 'string'
      earlyStoppingPatience: 'string'
      enableOnnxNormalization: 'string'
      evaluationFrequency: 'string'
      gradientAccumulationStep: 'string'
      imageSize: 'string'
      layersToFreeze: 'string'
      learningRate: 'string'
      learningRateScheduler: 'string'
      maxSize: 'string'
      minSize: 'string'
      modelName: 'string'
      modelSize: 'string'
      momentum: 'string'
      multiScale: 'string'
      nesterov: 'string'
      nmsIouThreshold: 'string'
      numberOfEpochs: 'string'
      numberOfWorkers: 'string'
      optimizer: 'string'
      randomSeed: 'string'
      stepLRGamma: 'string'
      stepLRStepSize: 'string'
      tileGridSize: 'string'
      tileOverlapRatio: 'string'
      tilePredictionsNmsThreshold: 'string'
      trainingBatchSize: 'string'
      validationBatchSize: 'string'
      validationIouThreshold: 'string'
      validationMetricType: 'string'
      warmupCosineLRCycles: 'string'
      warmupCosineLRWarmupEpochs: 'string'
      weightDecay: 'string'
    }
  ]
  sweepSettings: {
    earlyTermination: {
      delayEvaluation: int
      evaluationInterval: int
      policyType: 'string'
      // For remaining properties, see EarlyTerminationPolicy objects
    }
    samplingAlgorithm: 'string'
  }
  taskType: 'ImageObjectDetection'
  validationData: {
    description: 'string'
    jobInputType: 'string'
    mode: 'string'
    uri: 'string'
  }
  validationDataSize: int
}

For Regression, use:

{
  cvSplitColumnNames: [
    'string'
  ]
  featurizationSettings: {
    blockedTransformers: [
      'string'
    ]
    columnNameAndTypes: {
      {customized property}: 'string'
    }
    datasetLanguage: 'string'
    enableDnnFeaturization: bool
    mode: 'string'
    transformerParams: {
      {customized property}: [
        {
          fields: [
            'string'
          ]
          parameters: any(Azure.Bicep.Types.Concrete.AnyType)
        }
      ]
    }
  }
  limitSettings: {
    enableEarlyTermination: bool
    exitScore: int
    maxConcurrentTrials: int
    maxCoresPerTrial: int
    maxTrials: int
    timeout: 'string'
    trialTimeout: 'string'
  }
  nCrossValidations: {
    mode: 'string'
    // For remaining properties, see NCrossValidations objects
  }
  primaryMetric: 'string'
  taskType: 'Regression'
  testData: {
    description: 'string'
    jobInputType: 'string'
    mode: 'string'
    uri: 'string'
  }
  testDataSize: int
  trainingSettings: {
    allowedTrainingAlgorithms: [
      'string'
    ]
    blockedTrainingAlgorithms: [
      'string'
    ]
    enableDnnTraining: bool
    enableModelExplainability: bool
    enableOnnxCompatibleModels: bool
    enableStackEnsemble: bool
    enableVoteEnsemble: bool
    ensembleModelDownloadTimeout: 'string'
    stackEnsembleSettings: {
      stackMetaLearnerKWargs: any(Azure.Bicep.Types.Concrete.AnyType)
      stackMetaLearnerTrainPercentage: int
      stackMetaLearnerType: 'string'
    }
  }
  validationData: {
    description: 'string'
    jobInputType: 'string'
    mode: 'string'
    uri: 'string'
  }
  validationDataSize: int
  weightColumnName: 'string'
}

For TextClassification, use:

{
  featurizationSettings: {
    datasetLanguage: 'string'
  }
  limitSettings: {
    maxConcurrentTrials: int
    maxTrials: int
    timeout: 'string'
  }
  primaryMetric: 'string'
  taskType: 'TextClassification'
  validationData: {
    description: 'string'
    jobInputType: 'string'
    mode: 'string'
    uri: 'string'
  }
}

For TextClassificationMultilabel, use:

{
  featurizationSettings: {
    datasetLanguage: 'string'
  }
  limitSettings: {
    maxConcurrentTrials: int
    maxTrials: int
    timeout: 'string'
  }
  taskType: 'TextClassificationMultilabel'
  validationData: {
    description: 'string'
    jobInputType: 'string'
    mode: 'string'
    uri: 'string'
  }
}

For TextNER, use:

{
  featurizationSettings: {
    datasetLanguage: 'string'
  }
  limitSettings: {
    maxConcurrentTrials: int
    maxTrials: int
    timeout: 'string'
  }
  taskType: 'TextNER'
  validationData: {
    description: 'string'
    jobInputType: 'string'
    mode: 'string'
    uri: 'string'
  }
}

JobInput objects

Set the jobInputType property to specify the type of object.

For custom_model, use:

{
  jobInputType: 'custom_model'
  mode: 'string'
  uri: 'string'
}

For literal, use:

{
  jobInputType: 'literal'
  value: 'string'
}

For mlflow_model, use:

{
  jobInputType: 'mlflow_model'
  mode: 'string'
  uri: 'string'
}

For mltable, use:

{
  jobInputType: 'mltable'
  mode: 'string'
  uri: 'string'
}

For triton_model, use:

{
  jobInputType: 'triton_model'
  mode: 'string'
  uri: 'string'
}

For uri_file, use:

{
  jobInputType: 'uri_file'
  mode: 'string'
  uri: 'string'
}

For uri_folder, use:

{
  jobInputType: 'uri_folder'
  mode: 'string'
  uri: 'string'
}

Property values

AllNodes

Name	Description	Value
nodesValueType	[Required] Type of the Nodes value	'All' (required)

AmlToken

Name	Description	Value
identityType	[Required] Specifies the type of identity framework.	'AMLToken' (required)

AutoForecastHorizon

Name	Description	Value
mode	[Required] Set forecast horizon value selection mode.	'Auto' (required)

AutoMLJob

Name	Description	Value
environmentId	The ARM resource ID of the Environment specification for the job. This is optional value to provide, if not provided, AutoML will default this to Production AutoML curated environment version when running the job.	string
environmentVariables	Environment variables included in the job.	AutoMLJobEnvironmentVariables
jobType	[Required] Specifies the type of job.	'AutoML' (required)
outputs	Mapping of output data bindings used in the job.	AutoMLJobOutputs
queueSettings	Queue settings for the job	QueueSettings
resources	Compute Resource configuration for the job.	JobResourceConfiguration
taskDetails	[Required] This represents scenario which can be one of Tables/NLP/Image	AutoMLVertical (required)

AutoMLJobEnvironmentVariables

Name	Description	Value

AutoMLJobOutputs

Name	Description	Value

AutoMLVertical

Name	Description	Value
logVerbosity	Log verbosity for the job.	'Critical' 'Debug' 'Error' 'Info' 'NotSet' 'Warning'
targetColumnName	Target column name: This is prediction values column. Also known as label column name in context of classification tasks.	string
taskType	Set to 'Classification' for type Classification. Set to 'Forecasting' for type Forecasting. Set to 'ImageClassification' for type ImageClassification. Set to 'ImageClassificationMultilabel' for type ImageClassificationMultilabel. Set to 'ImageInstanceSegmentation' for type ImageInstanceSegmentation. Set to 'ImageObjectDetection' for type ImageObjectDetection. Set to 'Regression' for type Regression. Set to 'TextClassification' for type TextClassification. Set to 'TextClassificationMultilabel' for type TextClassificationMultilabel. Set to 'TextNER' for type TextNer.	'Classification' 'Forecasting' 'ImageClassification' 'ImageClassificationMultilabel' 'ImageInstanceSegmentation' 'ImageObjectDetection' 'Regression' 'TextClassification' 'TextClassificationMultilabel' 'TextNER' (required)
trainingData	[Required] Training data input.	MLTableJobInput (required)

AutoNCrossValidations

Name	Description	Value
mode	[Required] Mode for determining N-Cross validations.	'Auto' (required)

AutoSeasonality

Name	Description	Value
mode	[Required] Seasonality mode.	'Auto' (required)

AutoTargetLags

Name	Description	Value
mode	[Required] Set target lags mode - Auto/Custom	'Auto' (required)

AutoTargetRollingWindowSize

Name	Description	Value
mode	[Required] TargetRollingWindowSiz detection mode.	'Auto' (required)

AzureDevOpsWebhook

Name	Description	Value
webhookType	[Required] Specifies the type of service to send a callback	'AzureDevOps' (required)

BanditPolicy

Name	Description	Value
policyType	[Required] Name of policy configuration	'Bandit' (required)
slackAmount	Absolute distance allowed from the best performing run.	int
slackFactor	Ratio of the allowed distance from the best performing run.	int

BayesianSamplingAlgorithm

Name	Description	Value
samplingAlgorithmType	[Required] The algorithm used for generating hyperparameter values, along with configuration properties	'Bayesian' (required)

Classification

Name	Description	Value
cvSplitColumnNames	Columns to use for CVSplit data.	string[]
featurizationSettings	Featurization inputs needed for AutoML job.	TableVerticalFeaturizationSettings
limitSettings	Execution constraints for AutoMLJob.	TableVerticalLimitSettings
nCrossValidations	Number of cross validation folds to be applied on training dataset when validation dataset is not provided.	NCrossValidations
positiveLabel	Positive label for binary metrics calculation.	string
primaryMetric	Primary metric for the task.	'Accuracy' 'AUCWeighted' 'AveragePrecisionScoreWeighted' 'NormMacroRecall' 'PrecisionScoreWeighted'
taskType	[Required] Task type for AutoMLJob.	'Classification' (required)
testData	Test data input.	MLTableJobInput
testDataSize	The fraction of test dataset that needs to be set aside for validation purpose. Values between (0.0 , 1.0) Applied when validation dataset is not provided.	int
trainingSettings	Inputs for training phase for an AutoML Job.	ClassificationTrainingSettings
validationData	Validation data inputs.	MLTableJobInput
validationDataSize	The fraction of training dataset that needs to be set aside for validation purpose. Values between (0.0 , 1.0) Applied when validation dataset is not provided.	int
weightColumnName	The name of the sample weight column. Automated ML supports a weighted column as an input, causing rows in the data to be weighted up or down.	string

ClassificationTrainingSettings

Name	Description	Value
allowedTrainingAlgorithms	Allowed models for classification task.	String array containing any of: 'BernoulliNaiveBayes' 'DecisionTree' 'ExtremeRandomTrees' 'GradientBoosting' 'KNN' 'LightGBM' 'LinearSVM' 'LogisticRegression' 'MultinomialNaiveBayes' 'RandomForest' 'SGD' 'SVM' 'XGBoostClassifier'
blockedTrainingAlgorithms	Blocked models for classification task.	String array containing any of: 'BernoulliNaiveBayes' 'DecisionTree' 'ExtremeRandomTrees' 'GradientBoosting' 'KNN' 'LightGBM' 'LinearSVM' 'LogisticRegression' 'MultinomialNaiveBayes' 'RandomForest' 'SGD' 'SVM' 'XGBoostClassifier'
enableDnnTraining	Enable recommendation of DNN models.	bool
enableModelExplainability	Flag to turn on explainability on best model.	bool
enableOnnxCompatibleModels	Flag for enabling onnx compatible models.	bool
enableStackEnsemble	Enable stack ensemble run.	bool
enableVoteEnsemble	Enable voting ensemble run.	bool
ensembleModelDownloadTimeout	During VotingEnsemble and StackEnsemble model generation, multiple fitted models from the previous child runs are downloaded. Configure this parameter with a higher value than 300 secs, if more time is needed.	string
stackEnsembleSettings	Stack ensemble settings for stack ensemble run.	StackEnsembleSettings

ColumnTransformer

Name	Description	Value
fields	Fields to apply transformer logic on.	string[]
parameters	Different properties to be passed to transformer. Input expected is dictionary of key,value pairs in JSON format.	any

CommandJob

Name	Description	Value
codeId	ARM resource ID of the code asset.	string
command	[Required] The command to execute on startup of the job. eg. "python train.py"	string Constraints: Min length = 1 Pattern = `[a-zA-Z0-9_]` (required)
distribution	Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null.	DistributionConfiguration
environmentId	[Required] The ARM resource ID of the Environment specification for the job.	string Constraints: Min length = 1 Pattern = `[a-zA-Z0-9_]` (required)
environmentVariables	Environment variables included in the job.	CommandJobEnvironmentVariables
inputs	Mapping of input data bindings used in the job.	CommandJobInputs
jobType	[Required] Specifies the type of job.	'Command' (required)
limits	Command Job limit.	CommandJobLimits
outputs	Mapping of output data bindings used in the job.	CommandJobOutputs
queueSettings	Queue settings for the job	QueueSettings
resources	Compute Resource configuration for the job.	JobResourceConfiguration

CommandJobEnvironmentVariables

Name	Description	Value

CommandJobInputs

Name	Description	Value

CommandJobLimits

Name	Description	Value
jobLimitsType	[Required] JobLimit type.	'Command' 'Sweep' (required)
timeout	The max run duration in ISO 8601 format, after which the job will be cancelled. Only supports duration with precision as low as Seconds.	string

CommandJobOutputs

Name	Description	Value

CustomForecastHorizon

Name	Description	Value
mode	[Required] Set forecast horizon value selection mode.	'Custom' (required)
value	[Required] Forecast horizon value.	int (required)

CustomModelJobInput

Name	Description	Value
jobInputType	[Required] Specifies the type of job.	'custom_model' (required)
mode	Input Asset Delivery Mode.	'Direct' 'Download' 'EvalDownload' 'EvalMount' 'ReadOnlyMount' 'ReadWriteMount'
uri	[Required] Input Asset URI.	string Constraints: Min length = 1 Pattern = `[a-zA-Z0-9_]` (required)

CustomModelJobOutput

Name	Description	Value
jobOutputType	[Required] Specifies the type of job.	'custom_model' (required)
mode	Output Asset Delivery Mode.	'Direct' 'ReadWriteMount' 'Upload'
uri	Output Asset URI.	string

CustomNCrossValidations

Name	Description	Value
mode	[Required] Mode for determining N-Cross validations.	'Custom' (required)
value	[Required] N-Cross validations value.	int (required)

CustomSeasonality

Name	Description	Value
mode	[Required] Seasonality mode.	'Custom' (required)
value	[Required] Seasonality value.	int (required)

CustomTargetLags

Name	Description	Value
mode	[Required] Set target lags mode - Auto/Custom	'Custom' (required)
values	[Required] Set target lags values.	int[] (required)

CustomTargetRollingWindowSize

Name	Description	Value
mode	[Required] TargetRollingWindowSiz detection mode.	'Custom' (required)
value	[Required] TargetRollingWindowSize value.	int (required)

DistributionConfiguration

Name	Description	Value
distributionType	Set to 'Mpi' for type Mpi. Set to 'PyTorch' for type PyTorch. Set to 'TensorFlow' for type TensorFlow.	'Mpi' 'PyTorch' 'TensorFlow' (required)

EarlyTerminationPolicy

Name	Description	Value
delayEvaluation	Number of intervals by which to delay the first evaluation.	int
evaluationInterval	Interval (number of runs) between policy evaluations.	int
policyType	Set to 'Bandit' for type BanditPolicy. Set to 'MedianStopping' for type MedianStoppingPolicy. Set to 'TruncationSelection' for type TruncationSelectionPolicy.	'Bandit' 'MedianStopping' 'TruncationSelection' (required)

ForecastHorizon

Name	Description	Value
mode	Set to 'Auto' for type AutoForecastHorizon. Set to 'Custom' for type CustomForecastHorizon.	'Auto' 'Custom' (required)

Forecasting

Name	Description	Value
cvSplitColumnNames	Columns to use for CVSplit data.	string[]
featurizationSettings	Featurization inputs needed for AutoML job.	TableVerticalFeaturizationSettings
forecastingSettings	Forecasting task specific inputs.	ForecastingSettings
limitSettings	Execution constraints for AutoMLJob.	TableVerticalLimitSettings
nCrossValidations	Number of cross validation folds to be applied on training dataset when validation dataset is not provided.	NCrossValidations
primaryMetric	Primary metric for forecasting task.	'NormalizedMeanAbsoluteError' 'NormalizedRootMeanSquaredError' 'R2Score' 'SpearmanCorrelation'
taskType	[Required] Task type for AutoMLJob.	'Forecasting' (required)
testData	Test data input.	MLTableJobInput
testDataSize	The fraction of test dataset that needs to be set aside for validation purpose. Values between (0.0 , 1.0) Applied when validation dataset is not provided.	int
trainingSettings	Inputs for training phase for an AutoML Job.	ForecastingTrainingSettings
validationData	Validation data inputs.	MLTableJobInput
validationDataSize	The fraction of training dataset that needs to be set aside for validation purpose. Values between (0.0 , 1.0) Applied when validation dataset is not provided.	int
weightColumnName	The name of the sample weight column. Automated ML supports a weighted column as an input, causing rows in the data to be weighted up or down.	string

ForecastingSettings

Name	Description	Value
countryOrRegionForHolidays	Country or region for holidays for forecasting tasks. These should be ISO 3166 two-letter country/region codes, for example 'US' or 'GB'.	string
cvStepSize	Number of periods between the origin time of one CV fold and the next fold. For example, if `CVStepSize` = 3 for daily data, the origin time for each fold will be three days apart.	int
featureLags	Flag for generating lags for the numeric features with 'auto' or null.	'Auto' 'None'
forecastHorizon	The desired maximum forecast horizon in units of time-series frequency.	ForecastHorizon
frequency	When forecasting, this parameter represents the period with which the forecast is desired, for example daily, weekly, yearly, etc. The forecast frequency is dataset frequency by default.	string
seasonality	Set time series seasonality as an integer multiple of the series frequency. If seasonality is set to 'auto', it will be inferred.	Seasonality
shortSeriesHandlingConfig	The parameter defining how if AutoML should handle short time series.	'Auto' 'Drop' 'None' 'Pad'
targetAggregateFunction	The function to be used to aggregate the time series target column to conform to a user specified frequency. If the TargetAggregateFunction is set i.e. not 'None', but the freq parameter is not set, the error is raised. The possible target aggregation functions are: "sum", "max", "min" and "mean".	'Max' 'Mean' 'Min' 'None' 'Sum'
targetLags	The number of past periods to lag from the target column.	TargetLags
targetRollingWindowSize	The number of past periods used to create a rolling window average of the target column.	TargetRollingWindowSize
timeColumnName	The name of the time column. This parameter is required when forecasting to specify the datetime column in the input data used for building the time series and inferring its frequency.	string
timeSeriesIdColumnNames	The names of columns used to group a timeseries. It can be used to create multiple series. If grain is not defined, the data set is assumed to be one time-series. This parameter is used with task type forecasting.	string[]
useStl	Configure STL Decomposition of the time-series target column.	'None' 'Season' 'SeasonTrend'

ForecastingTrainingSettings

Name	Description	Value
allowedTrainingAlgorithms	Allowed models for forecasting task.	String array containing any of: 'Arimax' 'AutoArima' 'Average' 'DecisionTree' 'ElasticNet' 'ExponentialSmoothing' 'ExtremeRandomTrees' 'GradientBoosting' 'KNN' 'LassoLars' 'LightGBM' 'Naive' 'Prophet' 'RandomForest' 'SeasonalAverage' 'SeasonalNaive' 'SGD' 'TCNForecaster' 'XGBoostRegressor'
blockedTrainingAlgorithms	Blocked models for forecasting task.	String array containing any of: 'Arimax' 'AutoArima' 'Average' 'DecisionTree' 'ElasticNet' 'ExponentialSmoothing' 'ExtremeRandomTrees' 'GradientBoosting' 'KNN' 'LassoLars' 'LightGBM' 'Naive' 'Prophet' 'RandomForest' 'SeasonalAverage' 'SeasonalNaive' 'SGD' 'TCNForecaster' 'XGBoostRegressor'
enableDnnTraining	Enable recommendation of DNN models.	bool
enableModelExplainability	Flag to turn on explainability on best model.	bool
enableOnnxCompatibleModels	Flag for enabling onnx compatible models.	bool
enableStackEnsemble	Enable stack ensemble run.	bool
enableVoteEnsemble	Enable voting ensemble run.	bool
ensembleModelDownloadTimeout	During VotingEnsemble and StackEnsemble model generation, multiple fitted models from the previous child runs are downloaded. Configure this parameter with a higher value than 300 secs, if more time is needed.	string
stackEnsembleSettings	Stack ensemble settings for stack ensemble run.	StackEnsembleSettings

GridSamplingAlgorithm

Name	Description	Value
samplingAlgorithmType	[Required] The algorithm used for generating hyperparameter values, along with configuration properties	'Grid' (required)

IdentityConfiguration

Name	Description	Value
identityType	Set to 'AMLToken' for type AmlToken. Set to 'Managed' for type ManagedIdentity. Set to 'UserIdentity' for type UserIdentity.	'AMLToken' 'Managed' 'UserIdentity' (required)

ImageClassification

Name	Description	Value
limitSettings	[Required] Limit settings for the AutoML job.	ImageLimitSettings (required)
modelSettings	Settings used for training the model.	ImageModelSettingsClassification
primaryMetric	Primary metric to optimize for this task.	'Accuracy' 'AUCWeighted' 'AveragePrecisionScoreWeighted' 'NormMacroRecall' 'PrecisionScoreWeighted'
searchSpace	Search space for sampling different combinations of models and their hyperparameters.	ImageModelDistributionSettingsClassification[]
sweepSettings	Model sweeping and hyperparameter sweeping related settings.	ImageSweepSettings
taskType	[Required] Task type for AutoMLJob.	'ImageClassification' (required)
validationData	Validation data inputs.	MLTableJobInput
validationDataSize	The fraction of training dataset that needs to be set aside for validation purpose. Values between (0.0 , 1.0) Applied when validation dataset is not provided.	int

ImageClassificationMultilabel

Name	Description	Value
limitSettings	[Required] Limit settings for the AutoML job.	ImageLimitSettings (required)
modelSettings	Settings used for training the model.	ImageModelSettingsClassification
primaryMetric	Primary metric to optimize for this task.	'Accuracy' 'AUCWeighted' 'AveragePrecisionScoreWeighted' 'IOU' 'NormMacroRecall' 'PrecisionScoreWeighted'
searchSpace	Search space for sampling different combinations of models and their hyperparameters.	ImageModelDistributionSettingsClassification[]
sweepSettings	Model sweeping and hyperparameter sweeping related settings.	ImageSweepSettings
taskType	[Required] Task type for AutoMLJob.	'ImageClassificationMultilabel' (required)
validationData	Validation data inputs.	MLTableJobInput
validationDataSize	The fraction of training dataset that needs to be set aside for validation purpose. Values between (0.0 , 1.0) Applied when validation dataset is not provided.	int

ImageInstanceSegmentation

Name	Description	Value
limitSettings	[Required] Limit settings for the AutoML job.	ImageLimitSettings (required)
modelSettings	Settings used for training the model.	ImageModelSettingsObjectDetection
primaryMetric	Primary metric to optimize for this task.	'MeanAveragePrecision'
searchSpace	Search space for sampling different combinations of models and their hyperparameters.	ImageModelDistributionSettingsObjectDetection[]
sweepSettings	Model sweeping and hyperparameter sweeping related settings.	ImageSweepSettings
taskType	[Required] Task type for AutoMLJob.	'ImageInstanceSegmentation' (required)
validationData	Validation data inputs.	MLTableJobInput
validationDataSize	The fraction of training dataset that needs to be set aside for validation purpose. Values between (0.0 , 1.0) Applied when validation dataset is not provided.	int

ImageLimitSettings

Name	Description	Value
maxConcurrentTrials	Maximum number of concurrent AutoML iterations.	int
maxTrials	Maximum number of AutoML iterations.	int
timeout	AutoML job timeout.	string

ImageModelDistributionSettingsClassification

Name	Description	Value
amsGradient	Enable AMSGrad when optimizer is 'adam' or 'adamw'.	string
augmentations	Settings for using Augmentations.	string
beta1	Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].	string
beta2	Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].	string
distributed	Whether to use distributer training.	string
earlyStopping	Enable early stopping logic during training.	string
earlyStoppingDelay	Minimum number of epochs or validation evaluations to wait before primary metric improvement is tracked for early stopping. Must be a positive integer.	string
earlyStoppingPatience	Minimum number of epochs or validation evaluations with no primary metric improvement before the run is stopped. Must be a positive integer.	string
enableOnnxNormalization	Enable normalization when exporting ONNX model.	string
evaluationFrequency	Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.	string
gradientAccumulationStep	Gradient accumulation means running a configured number of "GradAccumulationStep" steps without updating the model weights while accumulating the gradients of those steps, and then using the accumulated gradients to compute the weight updates. Must be a positive integer.	string
layersToFreeze	Number of layers to freeze for the model. Must be a positive integer. For instance, passing 2 as value for 'seresnext' means freezing layer0 and layer1. For a full list of models supported and details on layer freeze, please see: /azure/machine-learning/how-to-auto-train-image-models.	string
learningRate	Initial learning rate. Must be a float in the range [0, 1].	string
learningRateScheduler	Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'.	string
modelName	Name of the model to use for training. For more information on the available models please visit the official documentation: /azure/machine-learning/how-to-auto-train-image-models.	string
momentum	Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].	string
nesterov	Enable nesterov when optimizer is 'sgd'.	string
numberOfEpochs	Number of training epochs. Must be a positive integer.	string
numberOfWorkers	Number of data loader workers. Must be a non-negative integer.	string
optimizer	Type of optimizer. Must be either 'sgd', 'adam', or 'adamw'.	string
randomSeed	Random seed to be used when using deterministic training.	string
stepLRGamma	Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].	string
stepLRStepSize	Value of step size when learning rate scheduler is 'step'. Must be a positive integer.	string
trainingBatchSize	Training batch size. Must be a positive integer.	string
trainingCropSize	Image crop size that is input to the neural network for the training dataset. Must be a positive integer.	string
validationBatchSize	Validation batch size. Must be a positive integer.	string
validationCropSize	Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.	string
validationResizeSize	Image size to which to resize before cropping for validation dataset. Must be a positive integer.	string
warmupCosineLRCycles	Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].	string
warmupCosineLRWarmupEpochs	Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.	string
weightDecay	Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].	string
weightedLoss	Weighted loss. The accepted values are 0 for no weighted loss. 1 for weighted loss with sqrt.(class_weights). 2 for weighted loss with class_weights. Must be 0 or 1 or 2.	string

ImageModelDistributionSettingsObjectDetection

Name	Description	Value
amsGradient	Enable AMSGrad when optimizer is 'adam' or 'adamw'.	string
augmentations	Settings for using Augmentations.	string
beta1	Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].	string
beta2	Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].	string
boxDetectionsPerImage	Maximum number of detections per image, for all classes. Must be a positive integer. Note: This settings is not supported for the 'yolov5' algorithm.	string
boxScoreThreshold	During inference, only return proposals with a classification score greater than BoxScoreThreshold. Must be a float in the range[0, 1].	string
distributed	Whether to use distributer training.	string
earlyStopping	Enable early stopping logic during training.	string
earlyStoppingDelay	Minimum number of epochs or validation evaluations to wait before primary metric improvement is tracked for early stopping. Must be a positive integer.	string
earlyStoppingPatience	Minimum number of epochs or validation evaluations with no primary metric improvement before the run is stopped. Must be a positive integer.	string
enableOnnxNormalization	Enable normalization when exporting ONNX model.	string
evaluationFrequency	Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.	string
gradientAccumulationStep	Gradient accumulation means running a configured number of "GradAccumulationStep" steps without updating the model weights while accumulating the gradients of those steps, and then using the accumulated gradients to compute the weight updates. Must be a positive integer.	string
imageSize	Image size for train and validation. Must be a positive integer. Note: The training run may get into CUDA OOM if the size is too big. Note: This settings is only supported for the 'yolov5' algorithm.	string
layersToFreeze	Number of layers to freeze for the model. Must be a positive integer. For instance, passing 2 as value for 'seresnext' means freezing layer0 and layer1. For a full list of models supported and details on layer freeze, please see: /azure/machine-learning/how-to-auto-train-image-models.	string
learningRate	Initial learning rate. Must be a float in the range [0, 1].	string
learningRateScheduler	Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'.	string
maxSize	Maximum size of the image to be rescaled before feeding it to the backbone. Must be a positive integer. Note: training run may get into CUDA OOM if the size is too big. Note: This settings is not supported for the 'yolov5' algorithm.	string
minSize	Minimum size of the image to be rescaled before feeding it to the backbone. Must be a positive integer. Note: training run may get into CUDA OOM if the size is too big. Note: This settings is not supported for the 'yolov5' algorithm.	string
modelName	Name of the model to use for training. For more information on the available models please visit the official documentation: /azure/machine-learning/how-to-auto-train-image-models.	string
modelSize	Model size. Must be 'small', 'medium', 'large', or 'xlarge'. Note: training run may get into CUDA OOM if the model size is too big. Note: This settings is only supported for the 'yolov5' algorithm.	string
momentum	Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].	string
multiScale	Enable multi-scale image by varying image size by +/- 50%. Note: training run may get into CUDA OOM if no sufficient GPU memory. Note: This settings is only supported for the 'yolov5' algorithm.	string
nesterov	Enable nesterov when optimizer is 'sgd'.	string
nmsIouThreshold	IOU threshold used during inference in NMS post processing. Must be float in the range [0, 1].	string
numberOfEpochs	Number of training epochs. Must be a positive integer.	string
numberOfWorkers	Number of data loader workers. Must be a non-negative integer.	string
optimizer	Type of optimizer. Must be either 'sgd', 'adam', or 'adamw'.	string
randomSeed	Random seed to be used when using deterministic training.	string
stepLRGamma	Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].	string
stepLRStepSize	Value of step size when learning rate scheduler is 'step'. Must be a positive integer.	string
tileGridSize	The grid size to use for tiling each image. Note: TileGridSize must not be None to enable small object detection logic. A string containing two integers in mxn format. Note: This settings is not supported for the 'yolov5' algorithm.	string
tileOverlapRatio	Overlap ratio between adjacent tiles in each dimension. Must be float in the range [0, 1). Note: This settings is not supported for the 'yolov5' algorithm.	string
tilePredictionsNmsThreshold	The IOU threshold to use to perform NMS while merging predictions from tiles and image. Used in validation/ inference. Must be float in the range [0, 1]. Note: This settings is not supported for the 'yolov5' algorithm. NMS: Non-maximum suppression	string
trainingBatchSize	Training batch size. Must be a positive integer.	string
validationBatchSize	Validation batch size. Must be a positive integer.	string
validationIouThreshold	IOU threshold to use when computing validation metric. Must be float in the range [0, 1].	string
validationMetricType	Metric computation method to use for validation metrics. Must be 'none', 'coco', 'voc', or 'coco_voc'.	string
warmupCosineLRCycles	Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].	string
warmupCosineLRWarmupEpochs	Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.	string
weightDecay	Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].	string

ImageModelSettingsClassification

Name	Description	Value
advancedSettings	Settings for advanced scenarios.	string
amsGradient	Enable AMSGrad when optimizer is 'adam' or 'adamw'.	bool
augmentations	Settings for using Augmentations.	string
beta1	Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].	int
beta2	Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].	int
checkpointFrequency	Frequency to store model checkpoints. Must be a positive integer.	int
checkpointModel	The pretrained checkpoint model for incremental training.	MLFlowModelJobInput
checkpointRunId	The id of a previous run that has a pretrained checkpoint for incremental training.	string
distributed	Whether to use distributed training.	bool
earlyStopping	Enable early stopping logic during training.	bool
earlyStoppingDelay	Minimum number of epochs or validation evaluations to wait before primary metric improvement is tracked for early stopping. Must be a positive integer.	int
earlyStoppingPatience	Minimum number of epochs or validation evaluations with no primary metric improvement before the run is stopped. Must be a positive integer.	int
enableOnnxNormalization	Enable normalization when exporting ONNX model.	bool
evaluationFrequency	Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.	int
gradientAccumulationStep	Gradient accumulation means running a configured number of "GradAccumulationStep" steps without updating the model weights while accumulating the gradients of those steps, and then using the accumulated gradients to compute the weight updates. Must be a positive integer.	int
layersToFreeze	Number of layers to freeze for the model. Must be a positive integer. For instance, passing 2 as value for 'seresnext' means freezing layer0 and layer1. For a full list of models supported and details on layer freeze, please see: /azure/machine-learning/how-to-auto-train-image-models.	int
learningRate	Initial learning rate. Must be a float in the range [0, 1].	int
learningRateScheduler	Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'.	'None' 'Step' 'WarmupCosine'
modelName	Name of the model to use for training. For more information on the available models please visit the official documentation: /azure/machine-learning/how-to-auto-train-image-models.	string
momentum	Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].	int
nesterov	Enable nesterov when optimizer is 'sgd'.	bool
numberOfEpochs	Number of training epochs. Must be a positive integer.	int
numberOfWorkers	Number of data loader workers. Must be a non-negative integer.	int
optimizer	Type of optimizer.	'Adam' 'Adamw' 'None' 'Sgd'
randomSeed	Random seed to be used when using deterministic training.	int
stepLRGamma	Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].	int
stepLRStepSize	Value of step size when learning rate scheduler is 'step'. Must be a positive integer.	int
trainingBatchSize	Training batch size. Must be a positive integer.	int
trainingCropSize	Image crop size that is input to the neural network for the training dataset. Must be a positive integer.	int
validationBatchSize	Validation batch size. Must be a positive integer.	int
validationCropSize	Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.	int
validationResizeSize	Image size to which to resize before cropping for validation dataset. Must be a positive integer.	int
warmupCosineLRCycles	Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].	int
warmupCosineLRWarmupEpochs	Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.	int
weightDecay	Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].	int
weightedLoss	Weighted loss. The accepted values are 0 for no weighted loss. 1 for weighted loss with sqrt.(class_weights). 2 for weighted loss with class_weights. Must be 0 or 1 or 2.	int

ImageModelSettingsObjectDetection

Name	Description	Value
advancedSettings	Settings for advanced scenarios.	string
amsGradient	Enable AMSGrad when optimizer is 'adam' or 'adamw'.	bool
augmentations	Settings for using Augmentations.	string
beta1	Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].	int
beta2	Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].	int
boxDetectionsPerImage	Maximum number of detections per image, for all classes. Must be a positive integer. Note: This settings is not supported for the 'yolov5' algorithm.	int
boxScoreThreshold	During inference, only return proposals with a classification score greater than BoxScoreThreshold. Must be a float in the range[0, 1].	int
checkpointFrequency	Frequency to store model checkpoints. Must be a positive integer.	int
checkpointModel	The pretrained checkpoint model for incremental training.	MLFlowModelJobInput
checkpointRunId	The id of a previous run that has a pretrained checkpoint for incremental training.	string
distributed	Whether to use distributed training.	bool
earlyStopping	Enable early stopping logic during training.	bool
earlyStoppingDelay	Minimum number of epochs or validation evaluations to wait before primary metric improvement is tracked for early stopping. Must be a positive integer.	int
earlyStoppingPatience	Minimum number of epochs or validation evaluations with no primary metric improvement before the run is stopped. Must be a positive integer.	int
enableOnnxNormalization	Enable normalization when exporting ONNX model.	bool
evaluationFrequency	Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.	int
gradientAccumulationStep	Gradient accumulation means running a configured number of "GradAccumulationStep" steps without updating the model weights while accumulating the gradients of those steps, and then using the accumulated gradients to compute the weight updates. Must be a positive integer.	int
imageSize	Image size for train and validation. Must be a positive integer. Note: The training run may get into CUDA OOM if the size is too big. Note: This settings is only supported for the 'yolov5' algorithm.	int
layersToFreeze	Number of layers to freeze for the model. Must be a positive integer. For instance, passing 2 as value for 'seresnext' means freezing layer0 and layer1. For a full list of models supported and details on layer freeze, please see: /azure/machine-learning/how-to-auto-train-image-models.	int
learningRate	Initial learning rate. Must be a float in the range [0, 1].	int
learningRateScheduler	Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'.	'None' 'Step' 'WarmupCosine'
maxSize	Maximum size of the image to be rescaled before feeding it to the backbone. Must be a positive integer. Note: training run may get into CUDA OOM if the size is too big. Note: This settings is not supported for the 'yolov5' algorithm.	int
minSize	Minimum size of the image to be rescaled before feeding it to the backbone. Must be a positive integer. Note: training run may get into CUDA OOM if the size is too big. Note: This settings is not supported for the 'yolov5' algorithm.	int
modelName	Name of the model to use for training. For more information on the available models please visit the official documentation: /azure/machine-learning/how-to-auto-train-image-models.	string
modelSize	Model size. Must be 'small', 'medium', 'large', or 'xlarge'. Note: training run may get into CUDA OOM if the model size is too big. Note: This settings is only supported for the 'yolov5' algorithm.	'ExtraLarge' 'Large' 'Medium' 'None' 'Small'
momentum	Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].	int
multiScale	Enable multi-scale image by varying image size by +/- 50%. Note: training run may get into CUDA OOM if no sufficient GPU memory. Note: This settings is only supported for the 'yolov5' algorithm.	bool
nesterov	Enable nesterov when optimizer is 'sgd'.	bool
nmsIouThreshold	IOU threshold used during inference in NMS post processing. Must be a float in the range [0, 1].	int
numberOfEpochs	Number of training epochs. Must be a positive integer.	int
numberOfWorkers	Number of data loader workers. Must be a non-negative integer.	int
optimizer	Type of optimizer.	'Adam' 'Adamw' 'None' 'Sgd'
randomSeed	Random seed to be used when using deterministic training.	int
stepLRGamma	Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].	int
stepLRStepSize	Value of step size when learning rate scheduler is 'step'. Must be a positive integer.	int
tileGridSize	The grid size to use for tiling each image. Note: TileGridSize must not be None to enable small object detection logic. A string containing two integers in mxn format. Note: This settings is not supported for the 'yolov5' algorithm.	string
tileOverlapRatio	Overlap ratio between adjacent tiles in each dimension. Must be float in the range [0, 1). Note: This settings is not supported for the 'yolov5' algorithm.	int
tilePredictionsNmsThreshold	The IOU threshold to use to perform NMS while merging predictions from tiles and image. Used in validation/ inference. Must be float in the range [0, 1]. Note: This settings is not supported for the 'yolov5' algorithm.	int
trainingBatchSize	Training batch size. Must be a positive integer.	int
validationBatchSize	Validation batch size. Must be a positive integer.	int
validationIouThreshold	IOU threshold to use when computing validation metric. Must be float in the range [0, 1].	int
validationMetricType	Metric computation method to use for validation metrics.	'Coco' 'CocoVoc' 'None' 'Voc'
warmupCosineLRCycles	Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].	int
warmupCosineLRWarmupEpochs	Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.	int
weightDecay	Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].	int

ImageObjectDetection

Name	Description	Value
limitSettings	[Required] Limit settings for the AutoML job.	ImageLimitSettings (required)
modelSettings	Settings used for training the model.	ImageModelSettingsObjectDetection
primaryMetric	Primary metric to optimize for this task.	'MeanAveragePrecision'
searchSpace	Search space for sampling different combinations of models and their hyperparameters.	ImageModelDistributionSettingsObjectDetection[]
sweepSettings	Model sweeping and hyperparameter sweeping related settings.	ImageSweepSettings
taskType	[Required] Task type for AutoMLJob.	'ImageObjectDetection' (required)
validationData	Validation data inputs.	MLTableJobInput
validationDataSize	The fraction of training dataset that needs to be set aside for validation purpose. Values between (0.0 , 1.0) Applied when validation dataset is not provided.	int

ImageSweepSettings

Name	Description	Value
earlyTermination	Type of early termination policy.	EarlyTerminationPolicy
samplingAlgorithm	[Required] Type of the hyperparameter sampling algorithms.	'Bayesian' 'Grid' 'Random' (required)

JobBaseProperties

Name	Description	Value
componentId	ARM resource ID of the component resource.	string
computeId	ARM resource ID of the compute resource.	string
description	The asset description text.	string
displayName	Display name of job.	string
experimentName	The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.	string
identity	Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null. Defaults to AmlToken if null.	IdentityConfiguration
isArchived	Is the asset archived?	bool
jobType	Set to 'AutoML' for type AutoMLJob. Set to 'Command' for type CommandJob. Set to 'Pipeline' for type PipelineJob. Set to 'Spark' for type SparkJob. Set to 'Sweep' for type SweepJob.	'AutoML' 'Command' 'Pipeline' 'Spark' 'Sweep' (required)
notificationSetting	Notification setting for the job	NotificationSetting
properties	The asset property dictionary.	ResourceBaseProperties
services	List of JobEndpoints. For local jobs, a job endpoint will have an endpoint value of FileStreamObject.	JobBaseServices
tags	Tag dictionary. Tags can be added, removed, and updated.	ResourceBaseTags

JobBaseServices

Name	Description	Value

JobInput

Name	Description	Value
description	Description for the input.	string
jobInputType	Set to 'custom_model' for type CustomModelJobInput. Set to 'literal' for type LiteralJobInput. Set to 'mlflow_model' for type MLFlowModelJobInput. Set to 'mltable' for type MLTableJobInput. Set to 'triton_model' for type TritonModelJobInput. Set to 'uri_file' for type UriFileJobInput. Set to 'uri_folder' for type UriFolderJobInput.	'custom_model' 'literal' 'mlflow_model' 'mltable' 'triton_model' 'uri_file' 'uri_folder' (required)

JobOutput

Name	Description	Value
description	Description for the output.	string
jobOutputType	Set to 'custom_model' for type CustomModelJobOutput. Set to 'mlflow_model' for type MLFlowModelJobOutput. Set to 'mltable' for type MLTableJobOutput. Set to 'triton_model' for type TritonModelJobOutput. Set to 'uri_file' for type UriFileJobOutput. Set to 'uri_folder' for type UriFolderJobOutput.	'custom_model' 'mlflow_model' 'mltable' 'triton_model' 'uri_file' 'uri_folder' (required)

JobResourceConfiguration

Name	Description	Value
dockerArgs	Extra arguments to pass to the Docker run command. This would override any parameters that have already been set by the system, or in this section. This parameter is only supported for Azure ML compute types.	string
instanceCount	Optional number of instances or nodes used by the compute target.	int
instanceType	Optional type of VM used as supported by the compute target.	string
properties	Additional properties bag.	ResourceConfigurationProperties
shmSize	Size of the docker container's shared memory block. This should be in the format of (number)(unit) where number as to be greater than 0 and the unit can be one of b(bytes), k(kilobytes), m(megabytes), or g(gigabytes).	string Constraints: Pattern = `\d+[bBkKmMgG]`

JobService

Name	Description	Value
endpoint	Url for endpoint.	string
jobServiceType	Endpoint type.	string
nodes	Nodes that user would like to start the service on. If Nodes is not set or set to null, the service will only be started on leader node.	Nodes
port	Port for endpoint.	int
properties	Additional properties to set on the endpoint.	JobServiceProperties

JobServiceProperties

Name	Description	Value

LiteralJobInput

Name	Description	Value
jobInputType	[Required] Specifies the type of job.	'literal' (required)
value	[Required] Literal value for the input.	string Constraints: Min length = 1 Pattern = `[a-zA-Z0-9_]` (required)

ManagedIdentity

Name	Description	Value
clientId	Specifies a user-assigned identity by client ID. For system-assigned, do not set this field.	string Constraints: Min length = 36 Max length = 36 Pattern = `^[0-9a-fA-F]{8}-([0-9a-fA-F]{4}-){3}[0-9a-fA-F]{12}$`
identityType	[Required] Specifies the type of identity framework.	'Managed' (required)
objectId	Specifies a user-assigned identity by object ID. For system-assigned, do not set this field.	string Constraints: Min length = 36 Max length = 36 Pattern = `^[0-9a-fA-F]{8}-([0-9a-fA-F]{4}-){3}[0-9a-fA-F]{12}$`
resourceId	Specifies a user-assigned identity by ARM resource ID. For system-assigned, do not set this field.	string

MedianStoppingPolicy

Name	Description	Value
policyType	[Required] Name of policy configuration	'MedianStopping' (required)