Schedules - Create Or Update
Create or update schedule.
PUT https://management.azure.com/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.MachineLearningServices/workspaces/{workspaceName}/schedules/{name}?api-version=2023-10-01
URI Parameters
Name | In | Required | Type | Description |
---|---|---|---|---|
name
|
path | True |
string |
Schedule name. Regex pattern: |
resource
|
path | True |
string |
The name of the resource group. The name is case insensitive. |
subscription
|
path | True |
string |
The ID of the target subscription. |
workspace
|
path | True |
string |
Name of Azure Machine Learning workspace. Regex pattern: |
api-version
|
query | True |
string |
The API version to use for this operation. |
Request Body
Name | Required | Type | Description |
---|---|---|---|
properties | True |
[Required] Additional attributes of the entity. |
Responses
Name | Type | Description |
---|---|---|
200 OK |
Create or update request is successful. |
|
201 Created |
Created Headers
|
|
Other Status Codes |
Error |
Examples
CreateOrUpdate Schedule.
Sample request
PUT https://management.azure.com/subscriptions/00000000-1111-2222-3333-444444444444/resourceGroups/test-rg/providers/Microsoft.MachineLearningServices/workspaces/my-aml-workspace/schedules/string?api-version=2023-10-01
{
"properties": {
"description": "string",
"tags": {
"string": "string"
},
"properties": {
"string": "string"
},
"displayName": "string",
"isEnabled": false,
"trigger": {
"endTime": "string",
"startTime": "string",
"timeZone": "string",
"triggerType": "Cron",
"expression": "string"
},
"action": {
"actionType": "InvokeBatchEndpoint",
"endpointInvocationDefinition": {
"9965593e-526f-4b89-bb36-761138cf2794": null
}
}
}
}
Sample response
{
"id": "string",
"name": "string",
"type": "string",
"properties": {
"description": "string",
"tags": {
"string": "string"
},
"properties": {
"string": "string"
},
"displayName": "string",
"isEnabled": false,
"trigger": {
"endTime": "string",
"startTime": "string",
"timeZone": "string",
"triggerType": "Cron",
"expression": "string"
},
"action": {
"actionType": "InvokeBatchEndpoint",
"endpointInvocationDefinition": {
"d77a9a9a-4bb5-4c0c-8a77-459be8b82b9f": null
}
},
"provisioningState": "Succeeded"
},
"systemData": {
"createdAt": "2020-01-01T12:34:56.999Z",
"createdBy": "string",
"createdByType": "Key",
"lastModifiedAt": "2020-01-01T12:34:56.999Z",
"lastModifiedBy": "string",
"lastModifiedByType": "Application"
}
}
{
"id": "string",
"name": "string",
"type": "string",
"properties": {
"description": "string",
"tags": {
"string": "string"
},
"properties": {
"string": "string"
},
"displayName": "string",
"isEnabled": false,
"trigger": {
"endTime": "string",
"startTime": "string",
"timeZone": "string",
"triggerType": "Cron",
"expression": "string"
},
"action": {
"actionType": "InvokeBatchEndpoint",
"endpointInvocationDefinition": {
"13ea51e0-ff28-49c3-a85d-9b5199eb14e5": null
}
},
"provisioningState": "Failed"
},
"systemData": {
"createdAt": "2020-01-01T12:34:56.999Z",
"createdBy": "string",
"createdByType": "Key",
"lastModifiedAt": "2020-01-01T12:34:56.999Z",
"lastModifiedBy": "string",
"lastModifiedByType": "User"
}
}
Definitions
Name | Description |
---|---|
All |
|
All |
All nodes means the service will be running on all of the nodes of the job |
Aml |
AML Token identity configuration. |
Aml |
AML token compute identity definition. |
Auto |
Forecast horizon determined automatically by system. |
Auto |
AutoMLJob class. Use this class for executing AutoML tasks like Classification/Regression etc. See TaskType enum for all the tasks supported. |
Auto |
N-Cross validations determined automatically. |
Auto |
|
Auto |
|
Auto |
Target lags rolling window determined automatically. |
Bandit |
Defines an early termination policy based on slack criteria, and a frequency and delay interval for evaluation |
Bayesian |
Defines a Sampling Algorithm that generates values based on previous values |
Blocked |
Enum for all classification models supported by AutoML. |
Categorical |
|
Categorical |
|
Categorical |
|
Categorical |
|
Categorical |
|
Categorical |
|
Classification |
Classification task in AutoML Table vertical. |
Classification |
Enum for all classification models supported by AutoML. |
Classification |
Primary metrics for classification multilabel tasks. |
Classification |
Primary metrics for classification tasks. |
Classification |
Classification Training related configuration. |
Command |
Command job definition. |
Command |
Command Job limit class. |
created |
The type of identity that created the resource. |
Create |
|
Cron |
|
Custom |
The desired maximum forecast horizon in units of time-series frequency. |
Custom |
|
Custom |
|
Custom |
|
Custom |
|
Custom |
N-Cross validations are specified by user. |
Custom |
|
Custom |
|
Custom |
|
Data |
|
Data |
|
Distribution |
Enum to determine the job distribution type. |
Early |
|
Endpoint |
|
Error |
The resource management error additional info. |
Error |
The error detail. |
Error |
Error response |
Feature |
|
Feature |
|
Feature |
|
Feature |
The mode of operation for computing feature importance. |
Feature |
|
Feature |
Flag for generating lags for the numeric features. |
Feature |
|
Featurization |
Featurization mode - determines data featurization mode. |
Fixed |
Fixed input data definition. |
Forecast |
Enum to determine forecast horizon selection mode. |
Forecasting |
Forecasting task in AutoML Table vertical. |
Forecasting |
Enum for all forecasting models supported by AutoML. |
Forecasting |
Primary metrics for Forecasting task. |
Forecasting |
Forecasting specific parameters. |
Forecasting |
Forecasting Training related configuration. |
Goal |
Defines supported metric goals for hyperparameter tuning |
Grid |
Defines a Sampling Algorithm that exhaustively generates every value combination in the space |
Identity |
Enum to determine identity framework. |
Image |
Image Classification. Multi-class image classification is used when an image is classified with only a single label from a set of classes - e.g. each image is classified as either an image of a 'cat' or a 'dog' or a 'duck'. |
Image |
Image Classification Multilabel. Multi-label image classification is used when an image could have one or more labels from a set of labels - e.g. an image could be labeled with both 'cat' and 'dog'. |
Image |
Image Instance Segmentation. Instance segmentation is used to identify objects in an image at the pixel level, drawing a polygon around each object in the image. |
Image |
Limit settings for the AutoML job. |
Image |
Distribution expressions to sweep over values of model settings. Some examples are:
|
Image |
Distribution expressions to sweep over values of model settings. Some examples are:
|
Image |
Settings used for training the model. For more information on the available settings please visit the official documentation: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models. |
Image |
Settings used for training the model. For more information on the available settings please visit the official documentation: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models. |
Image |
Image Object Detection. Object detection is used to identify objects in an image and locate each object with a bounding box e.g. locate all dogs and cats in an image and draw a bounding box around each. |
Image |
Model sweeping and hyperparameter sweeping related settings. |
Input |
Enum to determine the input data delivery mode. |
Instance |
Primary metrics for InstanceSegmentation tasks. |
Job |
Enum to determine the Job Input Type. |
Job |
|
Job |
Enum to determine the Job Output Type. |
Job |
|
Job |
|
Job |
Job endpoint definition |
Job |
The status of a job. |
Job |
Enum to determine the job tier. |
Job |
Enum to determine the type of job. |
Learning |
Learning rate scheduler enum. |
Literal |
Literal input type. |
Log |
Enum for setting log verbosity. |
Managed |
Managed compute identity definition. |
Managed |
Managed identity configuration. |
Managed |
Managed service identity (system assigned and/or user assigned identities) |
Managed |
Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed). |
Median |
Defines an early termination policy based on running averages of the primary metric of all runs |
MLFlow |
|
MLFlow |
|
MLTable |
|
MLTable |
|
Model |
Image model size. |
Model |
Model task type enum. |
Monitor |
Monitor compute identity type enum. |
Monitor |
Monitor compute type enum. |
Monitor |
|
Monitor |
|
Monitoring |
|
Monitoring |
|
Monitoring |
Monitoring input data type enum. |
Monitoring |
|
Monitoring |
|
Monitoring |
Monitoring target definition. |
Monitoring |
|
Monitor |
|
Monitor |
Monitor serverless spark compute definition. |
Mpi |
MPI distribution configuration. |
NCross |
Determines how N-Cross validations value is determined. |
Nlp |
|
Nlp |
Job execution constraints. |
Nodes |
The enumerated types for the nodes value |
Numerical |
|
Numerical |
|
Numerical |
|
Numerical |
|
Numerical |
|
Numerical |
|
Object |
Primary metrics for Image ObjectDetection task. |
Objective |
Optimization objective. |
Output |
Output data delivery mode enums. |
Pipeline |
Pipeline Job definition: defines generic to MFE attributes. |
Prediction |
|
Py |
PyTorch distribution configuration. |
Queue |
|
Random |
Defines a Sampling Algorithm that generates values randomly |
Random |
The specific type of random algorithm |
Recurrence |
Enum to describe the frequency of a recurrence schedule |
Recurrence |
|
Recurrence |
|
Regression |
Regression task in AutoML Table vertical. |
Regression |
Enum for all Regression models supported by AutoML. |
Regression |
Primary metrics for Regression task. |
Regression |
Regression Training related configuration. |
Rolling |
Rolling input data definition. |
Sampling |
|
Schedule |
Base definition of a schedule |
Schedule |
|
Schedule |
|
Schedule |
Azure Resource Manager resource envelope. |
Seasonality |
Forecasting seasonality mode. |
Short |
The parameter defining how if AutoML should handle short time series. |
Stack |
Advances setting to customize StackEnsemble run. |
Stack |
The meta-learner is a model trained on the output of the individual heterogeneous models. Default meta-learners are LogisticRegression for classification tasks (or LogisticRegressionCV if cross-validation is enabled) and ElasticNet for regression/forecasting tasks (or ElasticNetCV if cross-validation is enabled). This parameter can be one of the following strings: LogisticRegression, LogisticRegressionCV, LightGBMClassifier, ElasticNet, ElasticNetCV, LightGBMRegressor, or LinearRegression |
Static |
Static input data definition. |
Stochastic |
Stochastic optimizer for image models. |
Sweep |
Sweep job definition. |
Sweep |
Sweep Job limit class. |
system |
Metadata pertaining to creation and last modification of the resource. |
Table |
Featurization Configuration. |
Table |
Job execution constraints. |
Target |
Target aggregate function. |
Target |
Target lags selection modes. |
Target |
Target rolling windows size mode. |
Task |
AutoMLJob Task type. |
Tensor |
TensorFlow distribution configuration. |
Text |
Text Classification task in AutoML NLP vertical. NLP - Natural Language Processing. |
Text |
Text Classification Multilabel task in AutoML NLP vertical. NLP - Natural Language Processing. |
Text |
Text-NER task in AutoML NLP vertical. NER - Named Entity Recognition. NLP - Natural Language Processing. |
Top |
|
Trial |
Trial component definition. |
Trigger |
|
Triton |
|
Triton |
|
Truncation |
Defines an early termination policy that cancels a given percentage of runs at each evaluation interval. |
Uri |
|
Uri |
|
Uri |
|
Uri |
|
User |
User assigned identity properties |
User |
User identity configuration. |
Use |
Configure STL Decomposition of the time-series target column. |
Validation |
Metric computation method to use for validation metrics in image tasks. |
Week |
Enum of weekday |
AllFeatures
Name | Type | Description |
---|---|---|
filterType | string: |
[Required] Specifies the feature filter to leverage when selecting features to calculate metrics over. |
AllNodes
All nodes means the service will be running on all of the nodes of the job
Name | Type | Description |
---|---|---|
nodesValueType |
string:
All |
[Required] Type of the Nodes value |
AmlToken
AML Token identity configuration.
Name | Type | Description |
---|---|---|
identityType |
string:
AMLToken |
[Required] Specifies the type of identity framework. |
AmlTokenComputeIdentity
AML token compute identity definition.
Name | Type | Description |
---|---|---|
computeIdentityType | string: |
[Required] Specifies the type of identity to use within the monitoring jobs. |
AutoForecastHorizon
Forecast horizon determined automatically by system.
Name | Type | Description |
---|---|---|
mode |
string:
Auto |
[Required] Set forecast horizon value selection mode. |
AutoMLJob
AutoMLJob class. Use this class for executing AutoML tasks like Classification/Regression etc. See TaskType enum for all the tasks supported.
Name | Type | Default value | Description |
---|---|---|---|
componentId |
string |
ARM resource ID of the component resource. |
|
computeId |
string |
ARM resource ID of the compute resource. |
|
description |
string |
The asset description text. |
|
displayName |
string |
Display name of job. |
|
environmentId |
string |
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. |
|
environmentVariables |
object |
Environment variables included in the job. |
|
experimentName |
string |
Default |
The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment. |
identity | IdentityConfiguration: |
Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null. Defaults to AmlToken if null. |
|
isArchived |
boolean |
False |
Is the asset archived? |
jobType |
string:
AutoML |
[Required] Specifies the type of job. |
|
outputs |
object |
Mapping of output data bindings used in the job. |
|
properties |
object |
The asset property dictionary. |
|
queueSettings |
Queue settings for the job |
||
resources | {} |
Compute Resource configuration for the job. |
|
services |
<string,
Job |
List of JobEndpoints. For local jobs, a job endpoint will have an endpoint value of FileStreamObject. |
|
status |
Status of the job. |
||
tags |
object |
Tag dictionary. Tags can be added, removed, and updated. |
|
taskDetails | AutoMLVertical: |
[Required] This represents scenario which can be one of Tables/NLP/Image |
AutoNCrossValidations
N-Cross validations determined automatically.
Name | Type | Description |
---|---|---|
mode |
string:
Auto |
[Required] Mode for determining N-Cross validations. |
AutoSeasonality
Name | Type | Description |
---|---|---|
mode |
string:
Auto |
[Required] Seasonality mode. |
AutoTargetLags
Name | Type | Description |
---|---|---|
mode |
string:
Auto |
[Required] Set target lags mode - Auto/Custom |
AutoTargetRollingWindowSize
Target lags rolling window determined automatically.
Name | Type | Description |
---|---|---|
mode |
string:
Auto |
[Required] TargetRollingWindowSiz detection mode. |
BanditPolicy
Defines an early termination policy based on slack criteria, and a frequency and delay interval for evaluation
Name | Type | Default value | Description |
---|---|---|---|
delayEvaluation |
integer |
0 |
Number of intervals by which to delay the first evaluation. |
evaluationInterval |
integer |
0 |
Interval (number of runs) between policy evaluations. |
policyType |
string:
Bandit |
[Required] Name of policy configuration |
|
slackAmount |
number |
0 |
Absolute distance allowed from the best performing run. |
slackFactor |
number |
0 |
Ratio of the allowed distance from the best performing run. |
BayesianSamplingAlgorithm
Defines a Sampling Algorithm that generates values based on previous values
Name | Type | Description |
---|---|---|
samplingAlgorithmType |
string:
Bayesian |
[Required] The algorithm used for generating hyperparameter values, along with configuration properties |
BlockedTransformers
Enum for all classification models supported by AutoML.
Name | Type | Description |
---|---|---|
CatTargetEncoder |
string |
Target encoding for categorical data. |
CountVectorizer |
string |
Count Vectorizer converts a collection of text documents to a matrix of token counts. |
HashOneHotEncoder |
string |
Hashing One Hot Encoder can turn categorical variables into a limited number of new features. This is often used for high-cardinality categorical features. |
LabelEncoder |
string |
Label encoder converts labels/categorical variables in a numerical form. |
NaiveBayes |
string |
Naive Bayes is a classified that is used for classification of discrete features that are categorically distributed. |
OneHotEncoder |
string |
Ohe hot encoding creates a binary feature transformation. |
TextTargetEncoder |
string |
Target encoding for text data. |
TfIdf |
string |
Tf-Idf stands for, term-frequency times inverse document-frequency. This is a common term weighting scheme for identifying information from documents. |
WoETargetEncoder |
string |
Weight of Evidence encoding is a technique used to encode categorical variables. It uses the natural log of the P(1)/P(0) to create weights. |
WordEmbedding |
string |
Word embedding helps represents words or phrases as a vector, or a series of numbers. |
CategoricalDataDriftMetric
Name | Type | Description |
---|---|---|
JensenShannonDistance |
string |
The Jensen Shannon Distance (JSD) metric. |
PearsonsChiSquaredTest |
string |
The Pearsons Chi Squared Test metric. |
PopulationStabilityIndex |
string |
The Population Stability Index (PSI) metric. |
CategoricalDataDriftMetricThreshold
Name | Type | Description |
---|---|---|
dataType |
string:
Categorical |
[Required] Specifies the data type of the metric threshold. |
metric |
[Required] The categorical data drift metric to calculate. |
|
threshold |
The threshold value. If null, a default value will be set depending on the selected metric. |
CategoricalDataQualityMetric
Name | Type | Description |
---|---|---|
DataTypeErrorRate |
string |
Calculates the rate of data type errors. |
NullValueRate |
string |
Calculates the rate of null values. |
OutOfBoundsRate |
string |
Calculates the rate values are out of bounds. |
CategoricalDataQualityMetricThreshold
Name | Type | Description |
---|---|---|
dataType |
string:
Categorical |
[Required] Specifies the data type of the metric threshold. |
metric |
[Required] The categorical data quality metric to calculate. |
|
threshold |
The threshold value. If null, a default value will be set depending on the selected metric. |
CategoricalPredictionDriftMetric
Name | Type | Description |
---|---|---|
JensenShannonDistance |
string |
The Jensen Shannon Distance (JSD) metric. |
PearsonsChiSquaredTest |
string |
The Pearsons Chi Squared Test metric. |
PopulationStabilityIndex |
string |
The Population Stability Index (PSI) metric. |
CategoricalPredictionDriftMetricThreshold
Name | Type | Description |
---|---|---|
dataType |
string:
Categorical |
[Required] Specifies the data type of the metric threshold. |
metric |
[Required] The categorical prediction drift metric to calculate. |
|
threshold |
The threshold value. If null, a default value will be set depending on the selected metric. |
Classification
Classification task in AutoML Table vertical.
Name | Type | Default value | Description |
---|---|---|---|
cvSplitColumnNames |
string[] |
Columns to use for CVSplit data. |
|
featurizationSettings |
Featurization inputs needed for AutoML job. |
||
limitSettings |
Execution constraints for AutoMLJob. |
||
logVerbosity | Info |
Log verbosity for the job. |
|
nCrossValidations | NCrossValidations: |
Number of cross validation folds to be applied on training dataset when validation dataset is not provided. |
|
positiveLabel |
string |
Positive label for binary metrics calculation. |
|
primaryMetric | AUCWeighted |
Primary metric for the task. |
|
targetColumnName |
string |
Target column name: This is prediction values column. Also known as label column name in context of classification tasks. |
|
taskType | string: |
[Required] Task type for AutoMLJob. |
|
testData |
Test data input. |
||
testDataSize |
number |
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. |
|
trainingData |
[Required] Training data input. |
||
trainingSettings |
Inputs for training phase for an AutoML Job. |
||
validationData |
Validation data inputs. |
||
validationDataSize |
number |
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. |
|
weightColumnName |
string |
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. |
ClassificationModels
Enum for all classification models supported by AutoML.
Name | Type | Description |
---|---|---|
BernoulliNaiveBayes |
string |
Naive Bayes classifier for multivariate Bernoulli models. |
DecisionTree |
string |
Decision Trees are a non-parametric supervised learning method used for both classification and regression tasks. The goal is to create a model that predicts the value of a target variable by learning simple decision rules inferred from the data features. |
ExtremeRandomTrees |
string |
Extreme Trees is an ensemble machine learning algorithm that combines the predictions from many decision trees. It is related to the widely used random forest algorithm. |
GradientBoosting |
string |
The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
KNN |
string |
K-nearest neighbors (KNN) algorithm uses 'feature similarity' to predict the values of new datapoints which further means that the new data point will be assigned a value based on how closely it matches the points in the training set. |
LightGBM |
string |
LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
LinearSVM |
string |
A support vector machine (SVM) is a supervised machine learning model that uses classification algorithms for two-group classification problems. After giving an SVM model sets of labeled training data for each category, they're able to categorize new text. Linear SVM performs best when input data is linear, i.e., data can be easily classified by drawing the straight line between classified values on a plotted graph. |
LogisticRegression |
string |
Logistic regression is a fundamental classification technique. It belongs to the group of linear classifiers and is somewhat similar to polynomial and linear regression. Logistic regression is fast and relatively uncomplicated, and it's convenient for you to interpret the results. Although it's essentially a method for binary classification, it can also be applied to multiclass problems. |
MultinomialNaiveBayes |
string |
The multinomial Naive Bayes classifier is suitable for classification with discrete features (e.g., word counts for text classification). The multinomial distribution normally requires integer feature counts. However, in practice, fractional counts such as tf-idf may also work. |
RandomForest |
string |
Random forest is a supervised learning algorithm. The "forest" it builds, is an ensemble of decision trees, usually trained with the “bagging” method. The general idea of the bagging method is that a combination of learning models increases the overall result. |
SGD |
string |
SGD: Stochastic gradient descent is an optimization algorithm often used in machine learning applications to find the model parameters that correspond to the best fit between predicted and actual outputs. |
SVM |
string |
A support vector machine (SVM) is a supervised machine learning model that uses classification algorithms for two-group classification problems. After giving an SVM model sets of labeled training data for each category, they're able to categorize new text. |
XGBoostClassifier |
string |
XGBoost: Extreme Gradient Boosting Algorithm. This algorithm is used for structured data where target column values can be divided into distinct class values. |
ClassificationMultilabelPrimaryMetrics
Primary metrics for classification multilabel tasks.
Name | Type | Description |
---|---|---|
AUCWeighted |
string |
AUC is the Area under the curve. This metric represents arithmetic mean of the score for each class, weighted by the number of true instances in each class. |
Accuracy |
string |
Accuracy is the ratio of predictions that exactly match the true class labels. |
AveragePrecisionScoreWeighted |
string |
The arithmetic mean of the average precision score for each class, weighted by the number of true instances in each class. |
IOU |
string |
Intersection Over Union. Intersection of predictions divided by union of predictions. |
NormMacroRecall |
string |
Normalized macro recall is recall macro-averaged and normalized, so that random performance has a score of 0, and perfect performance has a score of 1. |
PrecisionScoreWeighted |
string |
The arithmetic mean of precision for each class, weighted by number of true instances in each class. |
ClassificationPrimaryMetrics
Primary metrics for classification tasks.
Name | Type | Description |
---|---|---|
AUCWeighted |
string |
AUC is the Area under the curve. This metric represents arithmetic mean of the score for each class, weighted by the number of true instances in each class. |
Accuracy |
string |
Accuracy is the ratio of predictions that exactly match the true class labels. |
AveragePrecisionScoreWeighted |
string |
The arithmetic mean of the average precision score for each class, weighted by the number of true instances in each class. |
NormMacroRecall |
string |
Normalized macro recall is recall macro-averaged and normalized, so that random performance has a score of 0, and perfect performance has a score of 1. |
PrecisionScoreWeighted |
string |
The arithmetic mean of precision for each class, weighted by number of true instances in each class. |
ClassificationTrainingSettings
Classification Training related configuration.
Name | Type | Default value | Description |
---|---|---|---|
allowedTrainingAlgorithms |
Allowed models for classification task. |
||
blockedTrainingAlgorithms |
Blocked models for classification task. |
||
enableDnnTraining |
boolean |
False |
Enable recommendation of DNN models. |
enableModelExplainability |
boolean |
True |
Flag to turn on explainability on best model. |
enableOnnxCompatibleModels |
boolean |
False |
Flag for enabling onnx compatible models. |
enableStackEnsemble |
boolean |
True |
Enable stack ensemble run. |
enableVoteEnsemble |
boolean |
True |
Enable voting ensemble run. |
ensembleModelDownloadTimeout |
string |
PT5M |
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. |
stackEnsembleSettings |
Stack ensemble settings for stack ensemble run. |
CommandJob
Command job definition.
Name | Type | Default value | Description |
---|---|---|---|
codeId |
string |
ARM resource ID of the code asset. |
|
command |
string |
[Required] The command to execute on startup of the job. eg. "python train.py" |
|
componentId |
string |
ARM resource ID of the component resource. |
|
computeId |
string |
ARM resource ID of the compute resource. |
|
description |
string |
The asset description text. |
|
displayName |
string |
Display name of job. |
|
distribution | DistributionConfiguration: |
Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null. |
|
environmentId |
string |
[Required] The ARM resource ID of the Environment specification for the job. |
|
environmentVariables |
object |
Environment variables included in the job. |
|
experimentName |
string |
Default |
The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment. |
identity | IdentityConfiguration: |
Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null. Defaults to AmlToken if null. |
|
inputs |
object |
Mapping of input data bindings used in the job. |
|
isArchived |
boolean |
False |
Is the asset archived? |
jobType |
string:
Command |
[Required] Specifies the type of job. |
|
limits |
Command Job limit. |
||
outputs |
object |
Mapping of output data bindings used in the job. |
|
parameters |
object |
Input parameters. |
|
properties |
object |
The asset property dictionary. |
|
queueSettings |
Queue settings for the job |
||
resources | {} |
Compute Resource configuration for the job. |
|
services |
<string,
Job |
List of JobEndpoints. For local jobs, a job endpoint will have an endpoint value of FileStreamObject. |
|
status |
Status of the job. |
||
tags |
object |
Tag dictionary. Tags can be added, removed, and updated. |
CommandJobLimits
Command Job limit class.
Name | Type | Description |
---|---|---|
jobLimitsType |
string:
Command |
[Required] JobLimit type. |
timeout |
string |
The max run duration in ISO 8601 format, after which the job will be cancelled. Only supports duration with precision as low as Seconds. |
createdByType
The type of identity that created the resource.
Name | Type | Description |
---|---|---|
Application |
string |
|
Key |
string |
|
ManagedIdentity |
string |
|
User |
string |
CreateMonitorAction
Name | Type | Description |
---|---|---|
actionType |
string:
Create |
[Required] Specifies the action type of the schedule |
monitorDefinition |
[Required] Defines the monitor. |
CronTrigger
Name | Type | Default value | Description |
---|---|---|---|
endTime |
string |
Specifies end time of schedule in ISO 8601, but without a UTC offset. Refer https://en.wikipedia.org/wiki/ISO_8601. Recommented format would be "2022-06-01T00:00:01" If not present, the schedule will run indefinitely |
|
expression |
string |
[Required] Specifies cron expression of schedule. The expression should follow NCronTab format. |
|
startTime |
string |
Specifies start time of schedule in ISO 8601 format, but without a UTC offset. |
|
timeZone |
string |
UTC |
Specifies time zone in which the schedule runs. TimeZone should follow Windows time zone format. Refer: https://docs.microsoft.com/en-us/windows-hardware/manufacture/desktop/default-time-zones?view=windows-11 |
triggerType |
string:
Cron |
[Required] |
CustomForecastHorizon
The desired maximum forecast horizon in units of time-series frequency.
Name | Type | Description |
---|---|---|
mode |
string:
Custom |
[Required] Set forecast horizon value selection mode. |
value |
integer |
[Required] Forecast horizon value. |
CustomMetricThreshold
Name | Type | Description |
---|---|---|
metric |
string |
[Required] The user-defined metric to calculate. |
threshold |
The threshold value. If null, a default value will be set depending on the selected metric. |
CustomModelJobInput
Name | Type | Default value | Description |
---|---|---|---|
description |
string |
Description for the input. |
|
jobInputType |
string:
custom_model |
[Required] Specifies the type of job. |
|
mode | ReadOnlyMount |
Input Asset Delivery Mode. |
|
uri |
string |
[Required] Input Asset URI. |
CustomModelJobOutput
Name | Type | Default value | Description |
---|---|---|---|
description |
string |
Description for the output. |
|
jobOutputType |
string:
custom_model |
[Required] Specifies the type of job. |
|
mode | ReadWriteMount |
Output Asset Delivery Mode. |
|
uri |
string |
Output Asset URI. |
CustomMonitoringSignal
Name | Type | Description |
---|---|---|
componentId |
string |
[Required] Reference to the component asset used to calculate the custom metrics. |
inputAssets |
object |
Monitoring assets to take as input. Key is the component input port name, value is the data asset. |
inputs |
object |
Extra component parameters to take as input. Key is the component literal input port name, value is the parameter value. |
metricThresholds |
[Required] A list of metrics to calculate and their associated thresholds. |
|
notificationTypes |
The current notification mode for this signal. |
|
properties |
object |
Property dictionary. Properties can be added, but not removed or altered. |
signalType |
string:
Custom |
[Required] Specifies the type of signal to monitor. |
CustomNCrossValidations
N-Cross validations are specified by user.
Name | Type | Description |
---|---|---|
mode |
string:
Custom |
[Required] Mode for determining N-Cross validations. |
value |
integer |
[Required] N-Cross validations value. |
CustomSeasonality
Name | Type | Description |
---|---|---|
mode |
string:
Custom |
[Required] Seasonality mode. |
value |
integer |
[Required] Seasonality value. |
CustomTargetLags
Name | Type | Description |
---|---|---|
mode |
string:
Custom |
[Required] Set target lags mode - Auto/Custom |
values |
integer[] |
[Required] Set target lags values. |
CustomTargetRollingWindowSize
Name | Type | Description |
---|---|---|
mode |
string:
Custom |
[Required] TargetRollingWindowSiz detection mode. |
value |
integer |
[Required] TargetRollingWindowSize value. |
DataDriftMonitoringSignal
Name | Type | Description |
---|---|---|
featureDataTypeOverride |
object |
A dictionary that maps feature names to their respective data types. |
featureImportanceSettings |
The settings for computing feature importance. |
|
features | MonitoringFeatureFilterBase: |
The feature filter which identifies which feature to calculate drift over. |
metricThresholds | DataDriftMetricThresholdBase[]: |
[Required] A list of metrics to calculate and their associated thresholds. |
notificationTypes |
The current notification mode for this signal. |
|
productionData | MonitoringInputDataBase: |
[Required] The data which drift will be calculated for. |
properties |
object |
Property dictionary. Properties can be added, but not removed or altered. |
referenceData | MonitoringInputDataBase: |
[Required] The data to calculate drift against. |
signalType |
string:
Data |
[Required] Specifies the type of signal to monitor. |
DataQualityMonitoringSignal
Name | Type | Description |
---|---|---|
featureDataTypeOverride |
object |
A dictionary that maps feature names to their respective data types. |
featureImportanceSettings |
The settings for computing feature importance. |
|
features | MonitoringFeatureFilterBase: |
The features to calculate drift over. |
metricThresholds | DataQualityMetricThresholdBase[]: |
[Required] A list of metrics to calculate and their associated thresholds. |
notificationTypes |
The current notification mode for this signal. |
|
productionData | MonitoringInputDataBase: |
[Required] The data produced by the production service which drift will be calculated for. |
properties |
object |
Property dictionary. Properties can be added, but not removed or altered. |
referenceData | MonitoringInputDataBase: |
[Required] The data to calculate drift against. |
signalType |
string:
Data |
[Required] Specifies the type of signal to monitor. |
DistributionType
Enum to determine the job distribution type.
Name | Type | Description |
---|---|---|
Mpi |
string |
|
PyTorch |
string |
|
TensorFlow |
string |
EarlyTerminationPolicyType
Name | Type | Description |
---|---|---|
Bandit |
string |
|
MedianStopping |
string |
|
TruncationSelection |
string |
EndpointScheduleAction
Name | Type | Description |
---|---|---|
actionType |
string:
Invoke |
[Required] Specifies the action type of the schedule |
endpointInvocationDefinition |
object |
[Required] Defines Schedule action definition details. |
ErrorAdditionalInfo
The resource management error additional info.
Name | Type | Description |
---|---|---|
info |
object |
The additional info. |
type |
string |
The additional info type. |
ErrorDetail
The error detail.
Name | Type | Description |
---|---|---|
additionalInfo |
The error additional info. |
|
code |
string |
The error code. |
details |
The error details. |
|
message |
string |
The error message. |
target |
string |
The error target. |
ErrorResponse
Error response
Name | Type | Description |
---|---|---|
error |
The error object. |
FeatureAttributionDriftMonitoringSignal
Name | Type | Description |
---|---|---|
featureDataTypeOverride |
object |
A dictionary that maps feature names to their respective data types. |
featureImportanceSettings |
[Required] The settings for computing feature importance. |
|
metricThreshold |
[Required] A list of metrics to calculate and their associated thresholds. |
|
notificationTypes |
The current notification mode for this signal. |
|
productionData | MonitoringInputDataBase[]: |
[Required] The data which drift will be calculated for. |
properties |
object |
Property dictionary. Properties can be added, but not removed or altered. |
referenceData | MonitoringInputDataBase: |
[Required] The data to calculate drift against. |
signalType |
string:
Feature |
[Required] Specifies the type of signal to monitor. |
FeatureAttributionMetric
Name | Type | Description |
---|---|---|
NormalizedDiscountedCumulativeGain |
string |
The Normalized Discounted Cumulative Gain metric. |
FeatureAttributionMetricThreshold
Name | Type | Description |
---|---|---|
metric |
[Required] The feature attribution metric to calculate. |
|
threshold |
The threshold value. If null, a default value will be set depending on the selected metric. |
FeatureImportanceMode
The mode of operation for computing feature importance.
Name | Type | Description |
---|---|---|
Disabled |
string |
Disables computing feature importance within a signal. |
Enabled |
string |
Enables computing feature importance within a signal. |
FeatureImportanceSettings
Name | Type | Default value | Description |
---|---|---|---|
mode | Disabled |
The mode of operation for computing feature importance. |
|
targetColumn |
string |
The name of the target column within the input data asset. |
FeatureLags
Flag for generating lags for the numeric features.
Name | Type | Description |
---|---|---|
Auto |
string |
System auto-generates feature lags. |
None |
string |
No feature lags generated. |
FeatureSubset
Name | Type | Description |
---|---|---|
features |
string[] |
[Required] The list of features to include. |
filterType | string: |
[Required] Specifies the feature filter to leverage when selecting features to calculate metrics over. |
FeaturizationMode
Featurization mode - determines data featurization mode.
Name | Type | Description |
---|---|---|
Auto |
string |
Auto mode, system performs featurization without any custom featurization inputs. |
Custom |
string |
Custom featurization. |
Off |
string |
Featurization off. 'Forecasting' task cannot use this value. |
FixedInputData
Fixed input data definition.
Name | Type | Description |
---|---|---|
columns |
object |
Mapping of column names to special uses. |
dataContext |
string |
The context metadata of the data source. |
inputDataType |
string:
Fixed |
[Required] Specifies the type of signal to monitor. |
jobInputType |
[Required] Specifies the type of job. |
|
uri |
string |
[Required] Input Asset URI. |
ForecastHorizonMode
Enum to determine forecast horizon selection mode.
Name | Type | Description |
---|---|---|
Auto |
string |
Forecast horizon to be determined automatically. |
Custom |
string |
Use the custom forecast horizon. |
Forecasting
Forecasting task in AutoML Table vertical.
Name | Type | Default value | Description |
---|---|---|---|
cvSplitColumnNames |
string[] |
Columns to use for CVSplit data. |
|
featurizationSettings |
Featurization inputs needed for AutoML job. |
||
forecastingSettings |
Forecasting task specific inputs. |
||
limitSettings |
Execution constraints for AutoMLJob. |
||
logVerbosity | Info |
Log verbosity for the job. |
|
nCrossValidations | NCrossValidations: |
Number of cross validation folds to be applied on training dataset when validation dataset is not provided. |
|
primaryMetric | NormalizedRootMeanSquaredError |
Primary metric for forecasting task. |
|
targetColumnName |
string |
Target column name: This is prediction values column. Also known as label column name in context of classification tasks. |
|
taskType | string: |
[Required] Task type for AutoMLJob. |
|
testData |
Test data input. |
||
testDataSize |
number |
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. |
|
trainingData |
[Required] Training data input. |
||
trainingSettings |
Inputs for training phase for an AutoML Job. |
||
validationData |
Validation data inputs. |
||
validationDataSize |
number |
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. |
|
weightColumnName |
string |
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. |
ForecastingModels
Enum for all forecasting models supported by AutoML.
Name | Type | Description |
---|---|---|
Arimax |
string |
An Autoregressive Integrated Moving Average with Explanatory Variable (ARIMAX) model can be viewed as a multiple regression model with one or more autoregressive (AR) terms and/or one or more moving average (MA) terms. This method is suitable for forecasting when data is stationary/non stationary, and multivariate with any type of data pattern, i.e., level/trend /seasonality/cyclicity. |
AutoArima |
string |
Auto-Autoregressive Integrated Moving Average (ARIMA) model uses time-series data and statistical analysis to interpret the data and make future predictions. This model aims to explain data by using time series data on its past values and uses linear regression to make predictions. |
Average |
string |
The Average forecasting model makes predictions by carrying forward the average of the target values for each time-series in the training data. |
DecisionTree |
string |
Decision Trees are a non-parametric supervised learning method used for both classification and regression tasks. The goal is to create a model that predicts the value of a target variable by learning simple decision rules inferred from the data features. |
ElasticNet |
string |
Elastic net is a popular type of regularized linear regression that combines two popular penalties, specifically the L1 and L2 penalty functions. |
ExponentialSmoothing |
string |
Exponential smoothing is a time series forecasting method for univariate data that can be extended to support data with a systematic trend or seasonal component. |
ExtremeRandomTrees |
string |
Extreme Trees is an ensemble machine learning algorithm that combines the predictions from many decision trees. It is related to the widely used random forest algorithm. |
GradientBoosting |
string |
The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
KNN |
string |
K-nearest neighbors (KNN) algorithm uses 'feature similarity' to predict the values of new datapoints which further means that the new data point will be assigned a value based on how closely it matches the points in the training set. |
LassoLars |
string |
Lasso model fit with Least Angle Regression a.k.a. Lars. It is a Linear Model trained with an L1 prior as regularizer. |
LightGBM |
string |
LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
Naive |
string |
The Naive forecasting model makes predictions by carrying forward the latest target value for each time-series in the training data. |
Prophet |
string |
Prophet is a procedure for forecasting time series data based on an additive model where non-linear trends are fit with yearly, weekly, and daily seasonality, plus holiday effects. It works best with time series that have strong seasonal effects and several seasons of historical data. Prophet is robust to missing data and shifts in the trend, and typically handles outliers well. |
RandomForest |
string |
Random forest is a supervised learning algorithm. The "forest" it builds, is an ensemble of decision trees, usually trained with the “bagging” method. The general idea of the bagging method is that a combination of learning models increases the overall result. |
SGD |
string |
SGD: Stochastic gradient descent is an optimization algorithm often used in machine learning applications to find the model parameters that correspond to the best fit between predicted and actual outputs. It's an inexact but powerful technique. |
SeasonalAverage |
string |
The Seasonal Average forecasting model makes predictions by carrying forward the average value of the latest season of data for each time-series in the training data. |
SeasonalNaive |
string |
The Seasonal Naive forecasting model makes predictions by carrying forward the latest season of target values for each time-series in the training data. |
TCNForecaster |
string |
TCNForecaster: Temporal Convolutional Networks Forecaster. //TODO: Ask forecasting team for brief intro. |
XGBoostRegressor |
string |
XGBoostRegressor: Extreme Gradient Boosting Regressor is a supervised machine learning model using ensemble of base learners. |
ForecastingPrimaryMetrics
Primary metrics for Forecasting task.
Name | Type | Description |
---|---|---|
NormalizedMeanAbsoluteError |
string |
The Normalized Mean Absolute Error (NMAE) is a validation metric to compare the Mean Absolute Error (MAE) of (time) series with different scales. |
NormalizedRootMeanSquaredError |
string |
The Normalized Root Mean Squared Error (NRMSE) the RMSE facilitates the comparison between models with different scales. |
R2Score |
string |
The R2 score is one of the performance evaluation measures for forecasting-based machine learning models. |
SpearmanCorrelation |
string |
The Spearman's rank coefficient of correlation is a non-parametric measure of rank correlation. |
ForecastingSettings
Forecasting specific parameters.
Name | Type | Default value | Description |
---|---|---|---|
countryOrRegionForHolidays |
string |
Country or region for holidays for forecasting tasks. These should be ISO 3166 two-letter country/region codes, for example 'US' or 'GB'. |
|
cvStepSize |
integer |
Number of periods between the origin time of one CV fold and the next fold. For
example, if |
|
featureLags | None |
Flag for generating lags for the numeric features with 'auto' or null. |
|
forecastHorizon | ForecastHorizon: | {"Mode": "Custom", "Value": 1} |
The desired maximum forecast horizon in units of time-series frequency. |
frequency |
string |
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. |
|
seasonality | Seasonality: | {"Mode": "Auto"} |
Set time series seasonality as an integer multiple of the series frequency. If seasonality is set to 'auto', it will be inferred. |
shortSeriesHandlingConfig | Auto |
The parameter defining how if AutoML should handle short time series. |
|
targetAggregateFunction | None |
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". |
|
targetLags | TargetLags: |
The number of past periods to lag from the target column. |
|
targetRollingWindowSize | TargetRollingWindowSize: |
The number of past periods used to create a rolling window average of the target column. |
|
timeColumnName |
string |
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. |
|
timeSeriesIdColumnNames |
string[] |
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. |
|
useStl | None |
Configure STL Decomposition of the time-series target column. |
ForecastingTrainingSettings
Forecasting Training related configuration.
Name | Type | Default value | Description |
---|---|---|---|
allowedTrainingAlgorithms |
Allowed models for forecasting task. |
||
blockedTrainingAlgorithms |
Blocked models for forecasting task. |
||
enableDnnTraining |
boolean |
False |
Enable recommendation of DNN models. |
enableModelExplainability |
boolean |
True |
Flag to turn on explainability on best model. |
enableOnnxCompatibleModels |
boolean |
False |
Flag for enabling onnx compatible models. |
enableStackEnsemble |
boolean |
True |
Enable stack ensemble run. |
enableVoteEnsemble |
boolean |
True |
Enable voting ensemble run. |
ensembleModelDownloadTimeout |
string |
PT5M |
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. |
stackEnsembleSettings |
Stack ensemble settings for stack ensemble run. |
Goal
Defines supported metric goals for hyperparameter tuning
Name | Type | Description |
---|---|---|
Maximize |
string |
|
Minimize |
string |
GridSamplingAlgorithm
Defines a Sampling Algorithm that exhaustively generates every value combination in the space
Name | Type | Description |
---|---|---|
samplingAlgorithmType |
string:
Grid |
[Required] The algorithm used for generating hyperparameter values, along with configuration properties |
IdentityConfigurationType
Enum to determine identity framework.
Name | Type | Description |
---|---|---|
AMLToken |
string |
|
Managed |
string |
|
UserIdentity |
string |
ImageClassification
Image Classification. Multi-class image classification is used when an image is classified with only a single label from a set of classes - e.g. each image is classified as either an image of a 'cat' or a 'dog' or a 'duck'.
Name | Type | Default value | Description |
---|---|---|---|
limitSettings |
[Required] Limit settings for the AutoML job. |
||
logVerbosity | Info |
Log verbosity for the job. |
|
modelSettings |
Settings used for training the model. |
||
primaryMetric | Accuracy |
Primary metric to optimize for this task. |
|
searchSpace |
Search space for sampling different combinations of models and their hyperparameters. |
||
sweepSettings |
Model sweeping and hyperparameter sweeping related settings. |
||
targetColumnName |
string |
Target column name: This is prediction values column. Also known as label column name in context of classification tasks. |
|
taskType | string: |
[Required] Task type for AutoMLJob. |
|
trainingData |
[Required] Training data input. |
||
validationData |
Validation data inputs. |
||
validationDataSize |
number |
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. |
ImageClassificationMultilabel
Image Classification Multilabel. Multi-label image classification is used when an image could have one or more labels from a set of labels - e.g. an image could be labeled with both 'cat' and 'dog'.
Name | Type | Default value | Description |
---|---|---|---|
limitSettings |
[Required] Limit settings for the AutoML job. |
||
logVerbosity | Info |
Log verbosity for the job. |
|
modelSettings |
Settings used for training the model. |
||
primaryMetric | IOU |
Primary metric to optimize for this task. |
|
searchSpace |
Search space for sampling different combinations of models and their hyperparameters. |
||
sweepSettings |
Model sweeping and hyperparameter sweeping related settings. |
||
targetColumnName |
string |
Target column name: This is prediction values column. Also known as label column name in context of classification tasks. |
|
taskType | string: |
[Required] Task type for AutoMLJob. |
|
trainingData |
[Required] Training data input. |
||
validationData |
Validation data inputs. |
||
validationDataSize |
number |
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. |
ImageInstanceSegmentation
Image Instance Segmentation. Instance segmentation is used to identify objects in an image at the pixel level, drawing a polygon around each object in the image.
Name | Type | Default value | Description |
---|---|---|---|
limitSettings |
[Required] Limit settings for the AutoML job. |
||
logVerbosity | Info |
Log verbosity for the job. |
|
modelSettings |
Settings used for training the model. |
||
primaryMetric | MeanAveragePrecision |
Primary metric to optimize for this task. |
|
searchSpace |
Search space for sampling different combinations of models and their hyperparameters. |
||
sweepSettings |
Model sweeping and hyperparameter sweeping related settings. |
||
targetColumnName |
string |
Target column name: This is prediction values column. Also known as label column name in context of classification tasks. |
|
taskType | string: |
[Required] Task type for AutoMLJob. |
|
trainingData |
[Required] Training data input. |
||
validationData |
Validation data inputs. |
||
validationDataSize |
number |
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. |
ImageLimitSettings
Limit settings for the AutoML job.
Name | Type | Default value | Description |
---|---|---|---|
maxConcurrentTrials |
integer |
1 |
Maximum number of concurrent AutoML iterations. |
maxTrials |
integer |
1 |
Maximum number of AutoML iterations. |
timeout |
string |
P7D |
AutoML job timeout. |
ImageModelDistributionSettingsClassification
Distribution expressions to sweep over values of model settings. Some examples are:
ModelName = "choice('seresnext', 'resnest50')";
LearningRate = "uniform(0.001, 0.01)";
LayersToFreeze = "choice(0, 2)";
```</example>
For more details on how to compose distribution expressions please check the documentation:
https://docs.microsoft.com/en-us/azure/machine-learning/how-to-tune-hyperparameters
For more information on the available settings please visit the official documentation:
https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models.
Name | Type | Description |
---|---|---|
amsGradient |
string |
Enable AMSGrad when optimizer is 'adam' or 'adamw'. |
augmentations |
string |
Settings for using Augmentations. |
beta1 |
string |
Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1]. |
beta2 |
string |
Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1]. |
distributed |
string |
Whether to use distributer training. |
earlyStopping |
string |
Enable early stopping logic during training. |
earlyStoppingDelay |
string |
Minimum number of epochs or validation evaluations to wait before primary metric improvement is tracked for early stopping. Must be a positive integer. |
earlyStoppingPatience |
string |
Minimum number of epochs or validation evaluations with no primary metric improvement before the run is stopped. Must be a positive integer. |
enableOnnxNormalization |
string |
Enable normalization when exporting ONNX model. |
evaluationFrequency |
string |
Frequency to evaluate validation dataset to get metric scores. Must be a positive integer. |
gradientAccumulationStep |
string |
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. |
layersToFreeze |
string |
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: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models. |
learningRate |
string |
Initial learning rate. Must be a float in the range [0, 1]. |
learningRateScheduler |
string |
Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'. |
modelName |
string |
Name of the model to use for training. For more information on the available models please visit the official documentation: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models. |
momentum |
string |
Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1]. |
nesterov |
string |
Enable nesterov when optimizer is 'sgd'. |
numberOfEpochs |
string |
Number of training epochs. Must be a positive integer. |
numberOfWorkers |
string |
Number of data loader workers. Must be a non-negative integer. |
optimizer |
string |
Type of optimizer. Must be either 'sgd', 'adam', or 'adamw'. |
randomSeed |
string |
Random seed to be used when using deterministic training. |
stepLRGamma |
string |
Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1]. |
stepLRStepSize |
string |
Value of step size when learning rate scheduler is 'step'. Must be a positive integer. |
trainingBatchSize |
string |
Training batch size. Must be a positive integer. |
trainingCropSize |
string |
Image crop size that is input to the neural network for the training dataset. Must be a positive integer. |
validationBatchSize |
string |
Validation batch size. Must be a positive integer. |
validationCropSize |
string |
Image crop size that is input to the neural network for the validation dataset. Must be a positive integer. |
validationResizeSize |
string |
Image size to which to resize before cropping for validation dataset. Must be a positive integer. |
warmupCosineLRCycles |
string |
Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1]. |
warmupCosineLRWarmupEpochs |
string |
Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer. |
weightDecay |
string |
Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1]. |
weightedLoss |
string |
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. |
ImageModelDistributionSettingsObjectDetection
Distribution expressions to sweep over values of model settings. Some examples are:
ModelName = "choice('seresnext', 'resnest50')";
LearningRate = "uniform(0.001, 0.01)";
LayersToFreeze = "choice(0, 2)";
```</example>
For more details on how to compose distribution expressions please check the documentation:
https://docs.microsoft.com/en-us/azure/machine-learning/how-to-tune-hyperparameters
For more information on the available settings please visit the official documentation:
https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models.
Name | Type | Description |
---|---|---|
amsGradient |
string |
Enable AMSGrad when optimizer is 'adam' or 'adamw'. |
augmentations |
string |
Settings for using Augmentations. |
beta1 |
string |
Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1]. |
beta2 |
string |
Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1]. |
boxDetectionsPerImage |
string |
Maximum number of detections per image, for all classes. Must be a positive integer. Note: This settings is not supported for the 'yolov5' algorithm. |
boxScoreThreshold |
string |
During inference, only return proposals with a classification score greater than BoxScoreThreshold. Must be a float in the range[0, 1]. |
distributed |
string |
Whether to use distributer training. |
earlyStopping |
string |
Enable early stopping logic during training. |
earlyStoppingDelay |
string |
Minimum number of epochs or validation evaluations to wait before primary metric improvement is tracked for early stopping. Must be a positive integer. |
earlyStoppingPatience |
string |
Minimum number of epochs or validation evaluations with no primary metric improvement before the run is stopped. Must be a positive integer. |
enableOnnxNormalization |
string |
Enable normalization when exporting ONNX model. |
evaluationFrequency |
string |
Frequency to evaluate validation dataset to get metric scores. Must be a positive integer. |
gradientAccumulationStep |
string |
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. |
imageSize |
string |
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. |
layersToFreeze |
string |
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: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models. |
learningRate |
string |
Initial learning rate. Must be a float in the range [0, 1]. |
learningRateScheduler |
string |
Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'. |
maxSize |
string |
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. |
minSize |
string |
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. |
modelName |
string |
Name of the model to use for training. For more information on the available models please visit the official documentation: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models. |
modelSize |
string |
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. |
momentum |
string |
Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1]. |
multiScale |
string |
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. |
nesterov |
string |
Enable nesterov when optimizer is 'sgd'. |
nmsIouThreshold |
string |
IOU threshold used during inference in NMS post processing. Must be float in the range [0, 1]. |
numberOfEpochs |
string |
Number of training epochs. Must be a positive integer. |
numberOfWorkers |
string |
Number of data loader workers. Must be a non-negative integer. |
optimizer |
string |
Type of optimizer. Must be either 'sgd', 'adam', or 'adamw'. |
randomSeed |
string |
Random seed to be used when using deterministic training. |
stepLRGamma |
string |
Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1]. |
stepLRStepSize |
string |
Value of step size when learning rate scheduler is 'step'. Must be a positive integer. |
tileGridSize |
string |
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. |
tileOverlapRatio |
string |
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. |
tilePredictionsNmsThreshold |
string |
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 |
trainingBatchSize |
string |
Training batch size. Must be a positive integer. |
validationBatchSize |
string |
Validation batch size. Must be a positive integer. |
validationIouThreshold |
string |
IOU threshold to use when computing validation metric. Must be float in the range [0, 1]. |
validationMetricType |
string |
Metric computation method to use for validation metrics. Must be 'none', 'coco', 'voc', or 'coco_voc'. |
warmupCosineLRCycles |
string |
Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1]. |
warmupCosineLRWarmupEpochs |
string |
Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer. |
weightDecay |
string |
Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1]. |
ImageModelSettingsClassification
Settings used for training the model. For more information on the available settings please visit the official documentation: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models.
Name | Type | Default value | Description |
---|---|---|---|
advancedSettings |
string |
Settings for advanced scenarios. |
|
amsGradient |
boolean |
Enable AMSGrad when optimizer is 'adam' or 'adamw'. |
|
augmentations |
string |
Settings for using Augmentations. |
|
beta1 |
number |
Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1]. |
|
beta2 |
number |
Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1]. |
|
checkpointFrequency |
integer |
Frequency to store model checkpoints. Must be a positive integer. |
|
checkpointModel |
The pretrained checkpoint model for incremental training. |
||
checkpointRunId |
string |
The id of a previous run that has a pretrained checkpoint for incremental training. |
|
distributed |
boolean |
Whether to use distributed training. |
|
earlyStopping |
boolean |
Enable early stopping logic during training. |
|
earlyStoppingDelay |
integer |
Minimum number of epochs or validation evaluations to wait before primary metric improvement is tracked for early stopping. Must be a positive integer. |
|
earlyStoppingPatience |
integer |
Minimum number of epochs or validation evaluations with no primary metric improvement before the run is stopped. Must be a positive integer. |
|
enableOnnxNormalization |
boolean |
Enable normalization when exporting ONNX model. |
|
evaluationFrequency |
integer |
Frequency to evaluate validation dataset to get metric scores. Must be a positive integer. |
|
gradientAccumulationStep |
integer |
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. |
|
layersToFreeze |
integer |
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: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models. |
|
learningRate |
number |
Initial learning rate. Must be a float in the range [0, 1]. |
|
learningRateScheduler | None |
Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'. |
|
modelName |
string |
Name of the model to use for training. For more information on the available models please visit the official documentation: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models. |
|
momentum |
number |
Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1]. |
|
nesterov |
boolean |
Enable nesterov when optimizer is 'sgd'. |
|
numberOfEpochs |
integer |
Number of training epochs. Must be a positive integer. |
|
numberOfWorkers |
integer |
Number of data loader workers. Must be a non-negative integer. |
|
optimizer | None |
Type of optimizer. |
|
randomSeed |
integer |
Random seed to be used when using deterministic training. |
|
stepLRGamma |
number |
Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1]. |
|
stepLRStepSize |
integer |
Value of step size when learning rate scheduler is 'step'. Must be a positive integer. |
|
trainingBatchSize |
integer |
Training batch size. Must be a positive integer. |
|
trainingCropSize |
integer |
Image crop size that is input to the neural network for the training dataset. Must be a positive integer. |
|
validationBatchSize |
integer |
Validation batch size. Must be a positive integer. |
|
validationCropSize |
integer |
Image crop size that is input to the neural network for the validation dataset. Must be a positive integer. |
|
validationResizeSize |
integer |
Image size to which to resize before cropping for validation dataset. Must be a positive integer. |
|
warmupCosineLRCycles |
number |
Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1]. |
|
warmupCosineLRWarmupEpochs |
integer |
Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer. |
|
weightDecay |
number |
Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1]. |
|
weightedLoss |
integer |
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. |
ImageModelSettingsObjectDetection
Settings used for training the model. For more information on the available settings please visit the official documentation: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models.
Name | Type | Default value | Description |
---|---|---|---|
advancedSettings |
string |
Settings for advanced scenarios. |
|
amsGradient |
boolean |
Enable AMSGrad when optimizer is 'adam' or 'adamw'. |
|
augmentations |
string |
Settings for using Augmentations. |
|
beta1 |
number |
Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1]. |
|
beta2 |
number |
Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1]. |
|
boxDetectionsPerImage |
integer |
Maximum number of detections per image, for all classes. Must be a positive integer. Note: This settings is not supported for the 'yolov5' algorithm. |
|
boxScoreThreshold |
number |
During inference, only return proposals with a classification score greater than BoxScoreThreshold. Must be a float in the range[0, 1]. |
|
checkpointFrequency |
integer |
Frequency to store model checkpoints. Must be a positive integer. |
|
checkpointModel |
The pretrained checkpoint model for incremental training. |
||
checkpointRunId |
string |
The id of a previous run that has a pretrained checkpoint for incremental training. |
|
distributed |
boolean |
Whether to use distributed training. |
|
earlyStopping |
boolean |
Enable early stopping logic during training. |
|
earlyStoppingDelay |
integer |
Minimum number of epochs or validation evaluations to wait before primary metric improvement is tracked for early stopping. Must be a positive integer. |
|
earlyStoppingPatience |
integer |
Minimum number of epochs or validation evaluations with no primary metric improvement before the run is stopped. Must be a positive integer. |
|
enableOnnxNormalization |
boolean |
Enable normalization when exporting ONNX model. |
|
evaluationFrequency |
integer |
Frequency to evaluate validation dataset to get metric scores. Must be a positive integer. |
|
gradientAccumulationStep |
integer |
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. |
|
imageSize |
integer |
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. |
|
layersToFreeze |
integer |
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: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models. |
|
learningRate |
number |
Initial learning rate. Must be a float in the range [0, 1]. |
|
learningRateScheduler | None |
Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'. |
|
maxSize |
integer |
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. |
|
minSize |
integer |
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. |
|
modelName |
string |
Name of the model to use for training. For more information on the available models please visit the official documentation: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models. |
|
modelSize | None |
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. |
|
momentum |
number |
Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1]. |
|
multiScale |
boolean |
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. |
|
nesterov |
boolean |
Enable nesterov when optimizer is 'sgd'. |
|
nmsIouThreshold |
number |
IOU threshold used during inference in NMS post processing. Must be a float in the range [0, 1]. |
|
numberOfEpochs |
integer |
Number of training epochs. Must be a positive integer. |
|
numberOfWorkers |
integer |
Number of data loader workers. Must be a non-negative integer. |
|
optimizer | None |
Type of optimizer. |
|
randomSeed |
integer |
Random seed to be used when using deterministic training. |
|
stepLRGamma |
number |
Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1]. |
|
stepLRStepSize |
integer |
Value of step size when learning rate scheduler is 'step'. Must be a positive integer. |
|
tileGridSize |
string |
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. |
|
tileOverlapRatio |
number |
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. |
|
tilePredictionsNmsThreshold |
number |
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. |
|
trainingBatchSize |
integer |
Training batch size. Must be a positive integer. |
|
validationBatchSize |
integer |
Validation batch size. Must be a positive integer. |
|
validationIouThreshold |
number |
IOU threshold to use when computing validation metric. Must be float in the range [0, 1]. |
|
validationMetricType | None |
Metric computation method to use for validation metrics. |
|
warmupCosineLRCycles |
number |
Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1]. |
|
warmupCosineLRWarmupEpochs |
integer |
Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer. |
|
weightDecay |
number |
Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1]. |
ImageObjectDetection
Image Object Detection. Object detection is used to identify objects in an image and locate each object with a bounding box e.g. locate all dogs and cats in an image and draw a bounding box around each.
Name | Type | Default value | Description |
---|---|---|---|
limitSettings |
[Required] Limit settings for the AutoML job. |
||
logVerbosity | Info |
Log verbosity for the job. |
|
modelSettings |
Settings used for training the model. |
||
primaryMetric | MeanAveragePrecision |
Primary metric to optimize for this task. |
|
searchSpace |
Search space for sampling different combinations of models and their hyperparameters. |
||
sweepSettings |
Model sweeping and hyperparameter sweeping related settings. |
||
targetColumnName |
string |
Target column name: This is prediction values column. Also known as label column name in context of classification tasks. |
|
taskType | string: |
[Required] Task type for AutoMLJob. |
|
trainingData |
[Required] Training data input. |
||
validationData |
Validation data inputs. |
||
validationDataSize |
number |
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. |
ImageSweepSettings
Model sweeping and hyperparameter sweeping related settings.
Name | Type | Description |
---|---|---|
earlyTermination | EarlyTerminationPolicy: |
Type of early termination policy. |
samplingAlgorithm |
[Required] Type of the hyperparameter sampling algorithms. |
InputDeliveryMode
Enum to determine the input data delivery mode.
Name | Type | Description |
---|---|---|
Direct |
string |
|
Download |
string |
|
EvalDownload |
string |
|
EvalMount |
string |
|
ReadOnlyMount |
string |
|
ReadWriteMount |
string |
InstanceSegmentationPrimaryMetrics
Primary metrics for InstanceSegmentation tasks.
Name | Type | Description |
---|---|---|
MeanAveragePrecision |
string |
Mean Average Precision (MAP) is the average of AP (Average Precision). AP is calculated for each class and averaged to get the MAP. |
JobInputType
Enum to determine the Job Input Type.
Name | Type | Description |
---|---|---|
custom_model |
string |
|
literal |
string |
|
mlflow_model |
string |
|
mltable |
string |
|
triton_model |
string |
|
uri_file |
string |
|
uri_folder |
string |
JobLimitsType
Name | Type | Description |
---|---|---|
Command |
string |
|
Sweep |
string |
JobOutputType
Enum to determine the Job Output Type.
Name | Type | Description |
---|---|---|
custom_model |
string |
|
mlflow_model |
string |
|
mltable |
string |
|
triton_model |
string |
|
uri_file |
string |
|
uri_folder |
string |
JobResourceConfiguration
Name | Type | Default value | Description |
---|---|---|---|
dockerArgs |
string |
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. |
|
instanceCount |
integer |
1 |
Optional number of instances or nodes used by the compute target. |
instanceType |
string |
Optional type of VM used as supported by the compute target. |
|
properties |
object |
Additional properties bag. |
|
shmSize |
string |
2g |
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). |
JobScheduleAction
Name | Type | Description |
---|---|---|
actionType |
string:
Create |
[Required] Specifies the action type of the schedule |
jobDefinition | JobBase: |
[Required] Defines Schedule action definition details. |
JobService
Job endpoint definition
Name | Type | Description |
---|---|---|
endpoint |
string |
Url for endpoint. |
errorMessage |
string |
Any error in the service. |
jobServiceType |
string |
Endpoint type. |
nodes | 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. |
port |
integer |
Port for endpoint. |
properties |
object |
Additional properties to set on the endpoint. |
status |
string |
Status of endpoint. |
JobStatus
The status of a job.
Name | Type | Description |
---|---|---|
CancelRequested |
string |
Cancellation has been requested for the job. |
Canceled |
string |
Following cancellation request, the job is now successfully canceled. |
Completed |
string |
Job completed successfully. This reflects that both the job itself and output collection states completed successfully |
Failed |
string |
Job failed. |
Finalizing |
string |
Job is completed in the target. It is in output collection state now. |
NotResponding |
string |
When heartbeat is enabled, if the run isn't updating any information to RunHistory then the run goes to NotResponding state. NotResponding is the only state that is exempt from strict transition orders. A run can go from NotResponding to any of the previous states. |
NotStarted |
string |
Run hasn't started yet. |
Paused |
string |
The job is paused by users. Some adjustment to labeling jobs can be made only in paused state. |
Preparing |
string |
The run environment is being prepared. |
Provisioning |
string |
(Not used currently) It will be used if ES is creating the compute target. |
Queued |
string |
The job is queued in the compute target. For example, in BatchAI the job is in queued state, while waiting for all required nodes to be ready. |
Running |
string |
The job started to run in the compute target. |
Starting |
string |
Run has started. The user has a run ID. |
Unknown |
string |
Default job status if not mapped to all other statuses |
JobTier
Enum to determine the job tier.
Name | Type | Description |
---|---|---|
Basic |
string |
|
Null |
string |
|
Premium |
string |
|
Spot |
string |
|
Standard |
string |
JobType
Enum to determine the type of job.
Name | Type | Description |
---|---|---|
AutoML |
string |
|
Command |
string |
|
Pipeline |
string |
|
Sweep |
string |
LearningRateScheduler
Learning rate scheduler enum.
Name | Type | Description |
---|---|---|
None |
string |
No learning rate scheduler selected. |
Step |
string |
Step learning rate scheduler. |
WarmupCosine |
string |
Cosine Annealing With Warmup. |
LiteralJobInput
Literal input type.
Name | Type | Description |
---|---|---|
description |
string |
Description for the input. |
jobInputType |
string:
literal |
[Required] Specifies the type of job. |
value |
string |
[Required] Literal value for the input. |
LogVerbosity
Enum for setting log verbosity.
Name | Type | Description |
---|---|---|
Critical |
string |
Only critical statements logged. |
Debug |
string |
Debug and above log statements logged. |
Error |
string |
Error and above log statements logged. |
Info |
string |
Info and above log statements logged. |
NotSet |
string |
No logs emitted. |
Warning |
string |
Warning and above log statements logged. |
ManagedComputeIdentity
Managed compute identity definition.
Name | Type | Description |
---|---|---|
computeIdentityType | string: |
[Required] Specifies the type of identity to use within the monitoring jobs. |
identity |
The identity which will be leveraged by the monitoring jobs. |
ManagedIdentity
Managed identity configuration.
Name | Type | Description |
---|---|---|
clientId |
string |
Specifies a user-assigned identity by client ID. For system-assigned, do not set this field. |
identityType |
string:
Managed |
[Required] Specifies the type of identity framework. |
objectId |
string |
Specifies a user-assigned identity by object ID. For system-assigned, do not set this field. |
resourceId |
string |
Specifies a user-assigned identity by ARM resource ID. For system-assigned, do not set this field. |
ManagedServiceIdentity
Managed service identity (system assigned and/or user assigned identities)
Name | Type | Description |
---|---|---|
principalId |
string |
The service principal ID of the system assigned identity. This property will only be provided for a system assigned identity. |
tenantId |
string |
The tenant ID of the system assigned identity. This property will only be provided for a system assigned identity. |
type |
Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed). |
|
userAssignedIdentities |
<string,
User |
User-Assigned Identities |
ManagedServiceIdentityType
Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed).
Name | Type | Description |
---|---|---|
None |
string |
|
SystemAssigned |
string |
|
SystemAssigned,UserAssigned |
string |
|
UserAssigned |
string |
MedianStoppingPolicy
Defines an early termination policy based on running averages of the primary metric of all runs
Name | Type | Default value | Description |
---|---|---|---|
delayEvaluation |
integer |
0 |
Number of intervals by which to delay the first evaluation. |
evaluationInterval |
integer |
0 |
Interval (number of runs) between policy evaluations. |
policyType |
string:
Median |
[Required] Name of policy configuration |
MLFlowModelJobInput
Name | Type | Default value | Description |
---|---|---|---|
description |
string |
Description for the input. |
|
jobInputType |
string:
mlflow_model |
[Required] Specifies the type of job. |
|
mode | ReadOnlyMount |
Input Asset Delivery Mode. |
|
uri |
string |
[Required] Input Asset URI. |
MLFlowModelJobOutput
Name | Type | Default value | Description |
---|---|---|---|
description |
string |
Description for the output. |
|
jobOutputType |
string:
mlflow_model |
[Required] Specifies the type of job. |
|
mode | ReadWriteMount |
Output Asset Delivery Mode. |
|
uri |
string |
Output Asset URI. |
MLTableJobInput
Name | Type | Default value | Description |
---|---|---|---|
description |
string |
Description for the input. |
|
jobInputType |
string:
mltable |
[Required] Specifies the type of job. |
|
mode | ReadOnlyMount |
Input Asset Delivery Mode. |
|
uri |
string |
[Required] Input Asset URI. |
MLTableJobOutput
Name | Type | Default value | Description |
---|---|---|---|
description |
string |
Description for the output. |
|
jobOutputType |
string:
mltable |
[Required] Specifies the type of job. |
|
mode | ReadWriteMount |
Output Asset Delivery Mode. |
|
uri |
string |
Output Asset URI. |
ModelSize
Image model size.
Name | Type | Description |
---|---|---|
ExtraLarge |
string |
Extra large size. |
Large |
string |
Large size. |
Medium |
string |
Medium size. |
None |
string |
No value selected. |
Small |
string |
Small size. |
ModelTaskType
Model task type enum.
Name | Type | Description |
---|---|---|
Classification |
string |
|
Regression |
string |
MonitorComputeIdentityType
Monitor compute identity type enum.
Name | Type | Description |
---|---|---|
AmlToken |
string |
Authenticates through user's AML token. |
ManagedIdentity |
string |
Authenticates through a user-provided managed identity. |
MonitorComputeType
Monitor compute type enum.
Name | Type | Description |
---|---|---|
ServerlessSpark |
string |
Serverless Spark compute. |
MonitorDefinition
Name | Type | Description |
---|---|---|
alertNotificationSettings |
The monitor's notification settings. |
|
computeConfiguration | MonitorComputeConfigurationBase: |
[Required] The ARM resource ID of the compute resource to run the monitoring job on. |
monitoringTarget |
The entities targeted by the monitor. |
|
signals |
object |
[Required] The signals to monitor. |
MonitorEmailNotificationSettings
Name | Type | Description |
---|---|---|
emails |
string[] |
The email recipient list which has a limitation of 499 characters in total. |
MonitoringFeatureDataType
Name | Type | Description |
---|---|---|
Categorical |
string |
Used for features of categorical data type. |
Numerical |
string |
Used for features of numerical data type. |
MonitoringFeatureFilterType
Name | Type | Description |
---|---|---|
AllFeatures |
string |
Includes all features. |
FeatureSubset |
string |
Includes a user-defined subset of features. |
TopNByAttribution |
string |
Only includes the top contributing features, measured by feature attribution. |
MonitoringInputDataType
Monitoring input data type enum.
Name | Type | Description |
---|---|---|
Fixed |
string |
An input data with tabular format which doesn't require preprocessing. |
Rolling |
string |
An input data which rolls relatively to the monitor's current run time. |
Static |
string |
An input data with a fixed window size. |
MonitoringNotificationType
Name | Type | Description |
---|---|---|
AmlNotification |
string |
Enables email notifications through AML notifications. |
MonitoringSignalType
Name | Type | Description |
---|---|---|
Custom |
string |
Tracks a custom signal provided by users. |
DataDrift |
string |
Tracks model input data distribution change, comparing against training data or past production data. |
DataQuality |
string |
Tracks model input data integrity. |
FeatureAttributionDrift |
string |
Tracks feature importance change in production, comparing against feature importance at training time. |
PredictionDrift |
string |
Tracks prediction result data distribution change, comparing against validation/test label data or past production data. |
MonitoringTarget
Monitoring target definition.
Name | Type | Description |
---|---|---|
deploymentId |
string |
Reference to the deployment asset targeted by this monitor. |
modelId |
string |
Reference to the model asset targeted by this monitor. |
taskType |
[Required] The machine learning task type of the monitored model. |
MonitoringThreshold
Name | Type | Description |
---|---|---|
value |
number |
The threshold value. If null, the set default is dependent on the metric type. |
MonitorNotificationSettings
Name | Type | Description |
---|---|---|
emailNotificationSettings |
The AML notification email settings. |
MonitorServerlessSparkCompute
Monitor serverless spark compute definition.
Name | Type | Description |
---|---|---|
computeIdentity | MonitorComputeIdentityBase: |
[Required] The identity scheme leveraged to by the spark jobs running on serverless Spark. |
computeType |
string:
Serverless |
[Required] Specifies the type of signal to monitor. |
instanceType |
string |
[Required] The instance type running the Spark job. |
runtimeVersion |
string |
[Required] The Spark runtime version. |
Mpi
MPI distribution configuration.
Name | Type | Description |
---|---|---|
distributionType | string: |
[Required] Specifies the type of distribution framework. |
processCountPerInstance |
integer |
Number of processes per MPI node. |
NCrossValidationsMode
Determines how N-Cross validations value is determined.
Name | Type | Description |
---|---|---|
Auto |
string |
Determine N-Cross validations value automatically. Supported only for 'Forecasting' AutoML task. |
Custom |
string |
Use custom N-Cross validations value. |
NlpVerticalFeaturizationSettings
Name | Type | Description |
---|---|---|
datasetLanguage |
string |
Dataset language, useful for the text data. |
NlpVerticalLimitSettings
Job execution constraints.
Name | Type | Default value | Description |
---|---|---|---|
maxConcurrentTrials |
integer |
1 |
Maximum Concurrent AutoML iterations. |
maxTrials |
integer |
1 |
Number of AutoML iterations. |
timeout |
string |
P7D |
AutoML job timeout. |
NodesValueType
The enumerated types for the nodes value
Name | Type | Description |
---|---|---|
All |
string |
NumericalDataDriftMetric
Name | Type | Description |
---|---|---|
JensenShannonDistance |
string |
The Jensen Shannon Distance (JSD) metric. |
NormalizedWassersteinDistance |
string |
The Normalized Wasserstein Distance metric. |
PopulationStabilityIndex |
string |
The Population Stability Index (PSI) metric. |
TwoSampleKolmogorovSmirnovTest |
string |
The Two Sample Kolmogorov-Smirnov Test (two-sample K–S) metric. |
NumericalDataDriftMetricThreshold
Name | Type | Description |
---|---|---|
dataType |
string:
Numerical |
[Required] Specifies the data type of the metric threshold. |
metric |
[Required] The numerical data drift metric to calculate. |
|
threshold |
The threshold value. If null, a default value will be set depending on the selected metric. |
NumericalDataQualityMetric
Name | Type | Description |
---|---|---|
DataTypeErrorRate |
string |
Calculates the rate of data type errors. |
NullValueRate |
string |
Calculates the rate of null values. |
OutOfBoundsRate |
string |
Calculates the rate values are out of bounds. |
NumericalDataQualityMetricThreshold
Name | Type | Description |
---|---|---|
dataType |
string:
Numerical |
[Required] Specifies the data type of the metric threshold. |
metric |
[Required] The numerical data quality metric to calculate. |
|
threshold |
The threshold value. If null, a default value will be set depending on the selected metric. |
NumericalPredictionDriftMetric
Name | Type | Description |
---|---|---|
JensenShannonDistance |
string |
The Jensen Shannon Distance (JSD) metric. |
NormalizedWassersteinDistance |
string |
The Normalized Wasserstein Distance metric. |
PopulationStabilityIndex |
string |
The Population Stability Index (PSI) metric. |
TwoSampleKolmogorovSmirnovTest |
string |
The Two Sample Kolmogorov-Smirnov Test (two-sample K–S) metric. |
NumericalPredictionDriftMetricThreshold
Name | Type | Description |
---|---|---|
dataType |
string:
Numerical |
[Required] Specifies the data type of the metric threshold. |
metric |
[Required] The numerical prediction drift metric to calculate. |
|
threshold |
The threshold value. If null, a default value will be set depending on the selected metric. |
ObjectDetectionPrimaryMetrics
Primary metrics for Image ObjectDetection task.
Name | Type | Description |
---|---|---|
MeanAveragePrecision |
string |
Mean Average Precision (MAP) is the average of AP (Average Precision). AP is calculated for each class and averaged to get the MAP. |
Objective
Optimization objective.
Name | Type | Description |
---|---|---|
goal |
[Required] Defines supported metric goals for hyperparameter tuning |
|
primaryMetric |
string |
[Required] Name of the metric to optimize. |
OutputDeliveryMode
Output data delivery mode enums.
Name | Type | Description |
---|---|---|
ReadWriteMount |
string |
|
Upload |
string |
PipelineJob
Pipeline Job definition: defines generic to MFE attributes.
Name | Type | Default value | Description |
---|---|---|---|
componentId |
string |
ARM resource ID of the component resource. |
|
computeId |
string |
ARM resource ID of the compute resource. |
|
description |
string |
The asset description text. |
|
displayName |
string |
Display name of job. |
|
experimentName |
string |
Default |
The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment. |
identity | IdentityConfiguration: |
Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null. Defaults to AmlToken if null. |
|
inputs |
object |
Inputs for the pipeline job. |
|
isArchived |
boolean |
False |
Is the asset archived? |
jobType |
string:
Pipeline |
[Required] Specifies the type of job. |
|
jobs |
object |
Jobs construct the Pipeline Job. |
|
outputs |
object |
Outputs for the pipeline job |
|
properties |
object |
The asset property dictionary. |
|
services |
<string,
Job |
List of JobEndpoints. For local jobs, a job endpoint will have an endpoint value of FileStreamObject. |
|
settings |
object |
Pipeline settings, for things like ContinueRunOnStepFailure etc. |
|
sourceJobId |
string |
ARM resource ID of source job. |
|
status |
Status of the job. |
||
tags |
object |
Tag dictionary. Tags can be added, removed, and updated. |
PredictionDriftMonitoringSignal
Name | Type | Description |
---|---|---|
featureDataTypeOverride |
object |
A dictionary that maps feature names to their respective data types. |
metricThresholds | PredictionDriftMetricThresholdBase[]: |
[Required] A list of metrics to calculate and their associated thresholds. |
notificationTypes |
The current notification mode for this signal. |
|
productionData | MonitoringInputDataBase: |
[Required] The data which drift will be calculated for. |
properties |
object |
Property dictionary. Properties can be added, but not removed or altered. |
referenceData | MonitoringInputDataBase: |
[Required] The data to calculate drift against. |
signalType |
string:
Prediction |
[Required] Specifies the type of signal to monitor. |
PyTorch
PyTorch distribution configuration.
Name | Type | Description |
---|---|---|
distributionType | string: |
[Required] Specifies the type of distribution framework. |
processCountPerInstance |
integer |
Number of processes per node. |
QueueSettings
Name | Type | Default value | Description |
---|---|---|---|
jobTier | Null |
Controls the compute job tier |
RandomSamplingAlgorithm
Defines a Sampling Algorithm that generates values randomly
Name | Type | Default value | Description |
---|---|---|---|
rule | Random |
The specific type of random algorithm |
|
samplingAlgorithmType |
string:
Random |
[Required] The algorithm used for generating hyperparameter values, along with configuration properties |
|
seed |
integer |
An optional integer to use as the seed for random number generation |
RandomSamplingAlgorithmRule
The specific type of random algorithm
Name | Type | Description |
---|---|---|
Random |
string |
|
Sobol |
string |
RecurrenceFrequency
Enum to describe the frequency of a recurrence schedule
Name | Type | Description |
---|---|---|
Day |
string |
Day frequency |
Hour |
string |
Hour frequency |
Minute |
string |
Minute frequency |
Month |
string |
Month frequency |
Week |
string |
Week frequency |
RecurrenceSchedule
Name | Type | Description |
---|---|---|
hours |
integer[] |
[Required] List of hours for the schedule. |
minutes |
integer[] |
[Required] List of minutes for the schedule. |
monthDays |
integer[] |
List of month days for the schedule |
weekDays |
Week |
List of days for the schedule. |
RecurrenceTrigger
Name | Type | Default value | Description |
---|---|---|---|
endTime |
string |
Specifies end time of schedule in ISO 8601, but without a UTC offset. Refer https://en.wikipedia.org/wiki/ISO_8601. Recommented format would be "2022-06-01T00:00:01" If not present, the schedule will run indefinitely |
|
frequency |
[Required] The frequency to trigger schedule. |
||
interval |
integer |
[Required] Specifies schedule interval in conjunction with frequency |
|
schedule |
The recurrence schedule. |
||
startTime |
string |
Specifies start time of schedule in ISO 8601 format, but without a UTC offset. |
|
timeZone |
string |
UTC |
Specifies time zone in which the schedule runs. TimeZone should follow Windows time zone format. Refer: https://docs.microsoft.com/en-us/windows-hardware/manufacture/desktop/default-time-zones?view=windows-11 |
triggerType |
string:
Recurrence |
[Required] |
Regression
Regression task in AutoML Table vertical.
Name | Type | Default value | Description |
---|---|---|---|
cvSplitColumnNames |
string[] |
Columns to use for CVSplit data. |
|
featurizationSettings |
Featurization inputs needed for AutoML job. |
||
limitSettings |
Execution constraints for AutoMLJob. |
||
logVerbosity | Info |
Log verbosity for the job. |
|
nCrossValidations | NCrossValidations: |
Number of cross validation folds to be applied on training dataset when validation dataset is not provided. |
|
primaryMetric | NormalizedRootMeanSquaredError |
Primary metric for regression task. |
|
targetColumnName |
string |
Target column name: This is prediction values column. Also known as label column name in context of classification tasks. |
|
taskType | string: |
[Required] Task type for AutoMLJob. |
|
testData |
Test data input. |
||
testDataSize |
number |
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. |
|
trainingData |
[Required] Training data input. |
||
trainingSettings |
Inputs for training phase for an AutoML Job. |
||
validationData |
Validation data inputs. |
||
validationDataSize |
number |
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. |
|
weightColumnName |
string |
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. |
RegressionModels
Enum for all Regression models supported by AutoML.
Name | Type | Description |
---|---|---|
DecisionTree |
string |
Decision Trees are a non-parametric supervised learning method used for both classification and regression tasks. The goal is to create a model that predicts the value of a target variable by learning simple decision rules inferred from the data features. |
ElasticNet |
string |
Elastic net is a popular type of regularized linear regression that combines two popular penalties, specifically the L1 and L2 penalty functions. |
ExtremeRandomTrees |
string |
Extreme Trees is an ensemble machine learning algorithm that combines the predictions from many decision trees. It is related to the widely used random forest algorithm. |
GradientBoosting |
string |
The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
KNN |
string |
K-nearest neighbors (KNN) algorithm uses 'feature similarity' to predict the values of new datapoints which further means that the new data point will be assigned a value based on how closely it matches the points in the training set. |
LassoLars |
string |
Lasso model fit with Least Angle Regression a.k.a. Lars. It is a Linear Model trained with an L1 prior as regularizer. |
LightGBM |
string |
LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
RandomForest |
string |
Random forest is a supervised learning algorithm. The "forest" it builds, is an ensemble of decision trees, usually trained with the “bagging” method. The general idea of the bagging method is that a combination of learning models increases the overall result. |
SGD |
string |
SGD: Stochastic gradient descent is an optimization algorithm often used in machine learning applications to find the model parameters that correspond to the best fit between predicted and actual outputs. It's an inexact but powerful technique. |
XGBoostRegressor |
string |
XGBoostRegressor: Extreme Gradient Boosting Regressor is a supervised machine learning model using ensemble of base learners. |
RegressionPrimaryMetrics
Primary metrics for Regression task.
Name | Type | Description |
---|---|---|
NormalizedMeanAbsoluteError |
string |
The Normalized Mean Absolute Error (NMAE) is a validation metric to compare the Mean Absolute Error (MAE) of (time) series with different scales. |
NormalizedRootMeanSquaredError |
string |
The Normalized Root Mean Squared Error (NRMSE) the RMSE facilitates the comparison between models with different scales. |
R2Score |
string |
The R2 score is one of the performance evaluation measures for forecasting-based machine learning models. |
SpearmanCorrelation |
string |
The Spearman's rank coefficient of correlation is a nonparametric measure of rank correlation. |
RegressionTrainingSettings
Regression Training related configuration.
Name | Type | Default value | Description |
---|---|---|---|
allowedTrainingAlgorithms |
Allowed models for regression task. |
||
blockedTrainingAlgorithms |
Blocked models for regression task. |
||
enableDnnTraining |
boolean |
False |
Enable recommendation of DNN models. |
enableModelExplainability |
boolean |
True |
Flag to turn on explainability on best model. |
enableOnnxCompatibleModels |
boolean |
False |
Flag for enabling onnx compatible models. |
enableStackEnsemble |
boolean |
True |
Enable stack ensemble run. |
enableVoteEnsemble |
boolean |
True |
Enable voting ensemble run. |
ensembleModelDownloadTimeout |
string |
PT5M |
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. |
stackEnsembleSettings |
Stack ensemble settings for stack ensemble run. |
RollingInputData
Rolling input data definition.
Name | Type | Description |
---|---|---|
columns |
object |
Mapping of column names to special uses. |
dataContext |
string |
The context metadata of the data source. |
inputDataType |
string:
Rolling |
[Required] Specifies the type of signal to monitor. |
jobInputType |
[Required] Specifies the type of job. |
|
preprocessingComponentId |
string |
Reference to the component asset used to preprocess the data. |
uri |
string |
[Required] Input Asset URI. |
windowOffset |
string |
[Required] The time offset between the end of the data window and the monitor's current run time. |
windowSize |
string |
[Required] The size of the rolling data window. |
SamplingAlgorithmType
Name | Type | Description |
---|---|---|
Bayesian |
string |
|
Grid |
string |
|
Random |
string |
Schedule
Base definition of a schedule
Name | Type | Default value | Description |
---|---|---|---|
action | ScheduleActionBase: |
[Required] Specifies the action of the schedule |
|
description |
string |
The asset description text. |
|
displayName |
string |
Display name of schedule. |
|
isEnabled |
boolean |
True |
Is the schedule enabled? |
properties |
object |
The asset property dictionary. |
|
provisioningState |
Provisioning state for the schedule. |
||
tags |
object |
Tag dictionary. Tags can be added, removed, and updated. |
|
trigger | TriggerBase: |
[Required] Specifies the trigger details |
ScheduleActionType
Name | Type | Description |
---|---|---|
CreateJob |
string |
|
CreateMonitor |
string |
|
InvokeBatchEndpoint |
string |
ScheduleProvisioningStatus
Name | Type | Description |
---|---|---|
Canceled |
string |
|
Creating |
string |
|
Deleting |
string |
|
Failed |
string |
|
Succeeded |
string |
|
Updating |
string |
ScheduleResource
Azure Resource Manager resource envelope.
Name | Type | Description |
---|---|---|
id |
string |
Fully qualified resource ID for the resource. Ex - /subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/{resourceProviderNamespace}/{resourceType}/{resourceName} |
name |
string |
The name of the resource |
properties |
[Required] Additional attributes of the entity. |
|
systemData |
Azure Resource Manager metadata containing createdBy and modifiedBy information. |
|
type |
string |
The type of the resource. E.g. "Microsoft.Compute/virtualMachines" or "Microsoft.Storage/storageAccounts" |
SeasonalityMode
Forecasting seasonality mode.
Name | Type | Description |
---|---|---|
Auto |
string |
Seasonality to be determined automatically. |
Custom |
string |
Use the custom seasonality value. |
ShortSeriesHandlingConfiguration
The parameter defining how if AutoML should handle short time series.
Name | Type | Description |
---|---|---|
Auto |
string |
Short series will be padded if there are no long series, otherwise short series will be dropped. |
Drop |
string |
All the short series will be dropped. |
None |
string |
Represents no/null value. |
Pad |
string |
All the short series will be padded. |
StackEnsembleSettings
Advances setting to customize StackEnsemble run.
Name | Type | Default value | Description |
---|---|---|---|
stackMetaLearnerKWargs |
object |
Optional parameters to pass to the initializer of the meta-learner. |
|
stackMetaLearnerTrainPercentage |
number |
0.2 |
Specifies the proportion of the training set (when choosing train and validation type of training) to be reserved for training the meta-learner. Default value is 0.2. |
stackMetaLearnerType | None |
The meta-learner is a model trained on the output of the individual heterogeneous models. |
StackMetaLearnerType
The meta-learner is a model trained on the output of the individual heterogeneous models. Default meta-learners are LogisticRegression for classification tasks (or LogisticRegressionCV if cross-validation is enabled) and ElasticNet for regression/forecasting tasks (or ElasticNetCV if cross-validation is enabled). This parameter can be one of the following strings: LogisticRegression, LogisticRegressionCV, LightGBMClassifier, ElasticNet, ElasticNetCV, LightGBMRegressor, or LinearRegression
Name | Type | Description |
---|---|---|
ElasticNet |
string |
Default meta-learners are LogisticRegression for regression task. |
ElasticNetCV |
string |
Default meta-learners are LogisticRegression for regression task when CV is on. |
LightGBMClassifier |
string |
|
LightGBMRegressor |
string |
|
LinearRegression |
string |
|
LogisticRegression |
string |
Default meta-learners are LogisticRegression for classification tasks. |
LogisticRegressionCV |
string |
Default meta-learners are LogisticRegression for classification task when CV is on. |
None |
string |
StaticInputData
Static input data definition.
Name | Type | Description |
---|---|---|
columns |
object |
Mapping of column names to special uses. |
dataContext |
string |
The context metadata of the data source. |
inputDataType |
string:
Static |
[Required] Specifies the type of signal to monitor. |
jobInputType |
[Required] Specifies the type of job. |
|
preprocessingComponentId |
string |
Reference to the component asset used to preprocess the data. |
uri |
string |
[Required] Input Asset URI. |
windowEnd |
string |
[Required] The end date of the data window. |
windowStart |
string |
[Required] The start date of the data window. |
StochasticOptimizer
Stochastic optimizer for image models.
Name | Type | Description |
---|---|---|
Adam |
string |
Adam is algorithm the optimizes stochastic objective functions based on adaptive estimates of moments |
Adamw |
string |
AdamW is a variant of the optimizer Adam that has an improved implementation of weight decay. |
None |
string |
No optimizer selected. |
Sgd |
string |
Stochastic Gradient Descent optimizer. |
SweepJob
Sweep job definition.
Name | Type | Default value | Description |
---|---|---|---|
componentId |
string |
ARM resource ID of the component resource. |
|
computeId |
string |
ARM resource ID of the compute resource. |
|
description |
string |
The asset description text. |
|
displayName |
string |
Display name of job. |
|
earlyTermination | EarlyTerminationPolicy: |
Early termination policies enable canceling poor-performing runs before they complete |
|
experimentName |
string |
Default |
The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment. |
identity | IdentityConfiguration: |
Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null. Defaults to AmlToken if null. |
|
inputs |
object |
Mapping of input data bindings used in the job. |
|
isArchived |
boolean |
False |
Is the asset archived? |
jobType |
string:
Sweep |
[Required] Specifies the type of job. |
|
limits | {} |
Sweep Job limit. |
|
objective |
[Required] Optimization objective. |
||
outputs |
object |
Mapping of output data bindings used in the job. |
|
properties |
object |
The asset property dictionary. |
|
queueSettings |
Queue settings for the job |
||
samplingAlgorithm | SamplingAlgorithm: |
[Required] The hyperparameter sampling algorithm |
|
searchSpace |
object |
[Required] A dictionary containing each parameter and its distribution. The dictionary key is the name of the parameter |
|
services |
<string,
Job |
List of JobEndpoints. For local jobs, a job endpoint will have an endpoint value of FileStreamObject. |
|
status |
Status of the job. |
||
tags |
object |
Tag dictionary. Tags can be added, removed, and updated. |
|
trial |
[Required] Trial component definition. |
SweepJobLimits
Sweep Job limit class.
Name | Type | Description |
---|---|---|
jobLimitsType |
string:
Sweep |
[Required] JobLimit type. |
maxConcurrentTrials |
integer |
Sweep Job max concurrent trials. |
maxTotalTrials |
integer |
Sweep Job max total trials. |
timeout |
string |
The max run duration in ISO 8601 format, after which the job will be cancelled. Only supports duration with precision as low as Seconds. |
trialTimeout |
string |
Sweep Job Trial timeout value. |
systemData
Metadata pertaining to creation and last modification of the resource.
Name | Type | Description |
---|---|---|
createdAt |
string |
The timestamp of resource creation (UTC). |
createdBy |
string |
The identity that created the resource. |
createdByType |
The type of identity that created the resource. |
|
lastModifiedAt |
string |
The timestamp of resource last modification (UTC) |
lastModifiedBy |
string |
The identity that last modified the resource. |
lastModifiedByType |
The type of identity that last modified the resource. |
TableVerticalFeaturizationSettings
Featurization Configuration.
Name | Type | Default value | Description |
---|---|---|---|
blockedTransformers |
These transformers shall not be used in featurization. |
||
columnNameAndTypes |
object |
Dictionary of column name and its type (int, float, string, datetime etc). |
|
datasetLanguage |
string |
Dataset language, useful for the text data. |
|
enableDnnFeaturization |
boolean |
False |
Determines whether to use Dnn based featurizers for data featurization. |
mode | Auto |
Featurization mode - User can keep the default 'Auto' mode and AutoML will take care of necessary transformation of the data in featurization phase. If 'Off' is selected then no featurization is done. If 'Custom' is selected then user can specify additional inputs to customize how featurization is done. |
|
transformerParams |
object |
User can specify additional transformers to be used along with the columns to which it would be applied and parameters for the transformer constructor. |
TableVerticalLimitSettings
Job execution constraints.
Name | Type | Default value | Description |
---|---|---|---|
enableEarlyTermination |
boolean |
True |
Enable early termination, determines whether or not if AutoMLJob will terminate early if there is no score improvement in last 20 iterations. |
exitScore |
number |
Exit score for the AutoML job. |
|
maxConcurrentTrials |
integer |
1 |
Maximum Concurrent iterations. |
maxCoresPerTrial |
integer |
-1 |
Max cores per iteration. |
maxTrials |
integer |
1000 |
Number of iterations. |
timeout |
string |
PT6H |
AutoML job timeout. |
trialTimeout |
string |
PT30M |
Iteration timeout. |
TargetAggregationFunction
Target aggregate function.
Name | Type | Description |
---|---|---|
Max |
string |
|
Mean |
string |
|
Min |
string |
|
None |
string |
Represent no value set. |
Sum |
string |
TargetLagsMode
Target lags selection modes.
Name | Type | Description |
---|---|---|
Auto |
string |
Target lags to be determined automatically. |
Custom |
string |
Use the custom target lags. |
TargetRollingWindowSizeMode
Target rolling windows size mode.
Name | Type | Description |
---|---|---|
Auto |
string |
Determine rolling windows size automatically. |
Custom |
string |
Use the specified rolling window size. |
TaskType
AutoMLJob Task type.
Name | Type | Description |
---|---|---|
Classification |
string |
Classification in machine learning and statistics is a supervised learning approach in which the computer program learns from the data given to it and make new observations or classifications. |
Forecasting |
string |
Forecasting is a special kind of regression task that deals with time-series data and creates forecasting model that can be used to predict the near future values based on the inputs. |
ImageClassification |
string |
Image Classification. Multi-class image classification is used when an image is classified with only a single label from a set of classes - e.g. each image is classified as either an image of a 'cat' or a 'dog' or a 'duck'. |
ImageClassificationMultilabel |
string |
Image Classification Multilabel. Multi-label image classification is used when an image could have one or more labels from a set of labels - e.g. an image could be labeled with both 'cat' and 'dog'. |
ImageInstanceSegmentation |
string |
Image Instance Segmentation. Instance segmentation is used to identify objects in an image at the pixel level, drawing a polygon around each object in the image. |
ImageObjectDetection |
string |
Image Object Detection. Object detection is used to identify objects in an image and locate each object with a bounding box e.g. locate all dogs and cats in an image and draw a bounding box around each. |
Regression |
string |
Regression means to predict the value using the input data. Regression models are used to predict a continuous value. |
TextClassification |
string |
Text classification (also known as text tagging or text categorization) is the process of sorting texts into categories. Categories are mutually exclusive. |
TextClassificationMultilabel |
string |
Multilabel classification task assigns each sample to a group (zero or more) of target labels. |
TextNER |
string |
Text Named Entity Recognition a.k.a. TextNER. Named Entity Recognition (NER) is the ability to take free-form text and identify the occurrences of entities such as people, locations, organizations, and more. |
TensorFlow
TensorFlow distribution configuration.
Name | Type | Default value | Description |
---|---|---|---|
distributionType | string: |
[Required] Specifies the type of distribution framework. |
|
parameterServerCount |
integer |
0 |
Number of parameter server tasks. |
workerCount |
integer |
Number of workers. If not specified, will default to the instance count. |
TextClassification
Text Classification task in AutoML NLP vertical. NLP - Natural Language Processing.
Name | Type | Default value | Description |
---|---|---|---|
featurizationSettings |
Featurization inputs needed for AutoML job. |
||
limitSettings |
Execution constraints for AutoMLJob. |
||
logVerbosity | Info |
Log verbosity for the job. |
|
primaryMetric | Accuracy |
Primary metric for Text-Classification task. |
|
targetColumnName |
string |
Target column name: This is prediction values column. Also known as label column name in context of classification tasks. |
|
taskType | string: |
[Required] Task type for AutoMLJob. |
|
trainingData |
[Required] Training data input. |
||
validationData |
Validation data inputs. |
TextClassificationMultilabel
Text Classification Multilabel task in AutoML NLP vertical. NLP - Natural Language Processing.
Name | Type | Default value | Description |
---|---|---|---|
featurizationSettings |
Featurization inputs needed for AutoML job. |
||
limitSettings |
Execution constraints for AutoMLJob. |
||
logVerbosity | Info |
Log verbosity for the job. |
|
primaryMetric |
Primary metric for Text-Classification-Multilabel task. Currently only Accuracy is supported as primary metric, hence user need not set it explicitly. |
||
targetColumnName |
string |
Target column name: This is prediction values column. Also known as label column name in context of classification tasks. |
|
taskType | string: |
[Required] Task type for AutoMLJob. |
|
trainingData |
[Required] Training data input. |
||
validationData |
Validation data inputs. |
TextNer
Text-NER task in AutoML NLP vertical. NER - Named Entity Recognition. NLP - Natural Language Processing.
Name | Type | Default value | Description |
---|---|---|---|
featurizationSettings |
Featurization inputs needed for AutoML job. |
||
limitSettings |
Execution constraints for AutoMLJob. |
||
logVerbosity | Info |
Log verbosity for the job. |
|
primaryMetric |
Primary metric for Text-NER task. Only 'Accuracy' is supported for Text-NER, so user need not set this explicitly. |
||
targetColumnName |
string |
Target column name: This is prediction values column. Also known as label column name in context of classification tasks. |
|
taskType |
string:
TextNER |
[Required] Task type for AutoMLJob. |
|
trainingData |
[Required] Training data input. |
||
validationData |
Validation data inputs. |
TopNFeaturesByAttribution
Name | Type | Default value | Description |
---|---|---|---|
filterType |
string:
Top |
[Required] Specifies the feature filter to leverage when selecting features to calculate metrics over. |
|
top |
integer |
10 |
The number of top features to include. |
TrialComponent
Trial component definition.
Name | Type | Default value | Description |
---|---|---|---|
codeId |
string |
ARM resource ID of the code asset. |
|
command |
string |
[Required] The command to execute on startup of the job. eg. "python train.py" |
|
distribution | DistributionConfiguration: |
Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null. |
|
environmentId |
string |
[Required] The ARM resource ID of the Environment specification for the job. |
|
environmentVariables |
object |
Environment variables included in the job. |
|
resources | {} |
Compute Resource configuration for the job. |
TriggerType
Name | Type | Description |
---|---|---|
Cron |
string |
|
Recurrence |
string |
TritonModelJobInput
Name | Type | Default value | Description |
---|---|---|---|
description |
string |
Description for the input. |
|
jobInputType |
string:
triton_model |
[Required] Specifies the type of job. |
|
mode | ReadOnlyMount |
Input Asset Delivery Mode. |
|
uri |
string |
[Required] Input Asset URI. |
TritonModelJobOutput
Name | Type | Default value | Description |
---|---|---|---|
description |
string |
Description for the output. |
|
jobOutputType |
string:
triton_model |
[Required] Specifies the type of job. |
|
mode | ReadWriteMount |
Output Asset Delivery Mode. |
|
uri |
string |
Output Asset URI. |
TruncationSelectionPolicy
Defines an early termination policy that cancels a given percentage of runs at each evaluation interval.
Name | Type | Default value | Description |
---|---|---|---|
delayEvaluation |
integer |
0 |
Number of intervals by which to delay the first evaluation. |
evaluationInterval |
integer |
0 |
Interval (number of runs) between policy evaluations. |
policyType |
string:
Truncation |
[Required] Name of policy configuration |
|
truncationPercentage |
integer |
0 |
The percentage of runs to cancel at each evaluation interval. |
UriFileJobInput
Name | Type | Default value | Description |
---|---|---|---|
description |
string |
Description for the input. |
|
jobInputType |
string:
uri_file |
[Required] Specifies the type of job. |
|
mode | ReadOnlyMount |
Input Asset Delivery Mode. |
|
uri |
string |
[Required] Input Asset URI. |
UriFileJobOutput
Name | Type | Default value | Description |
---|---|---|---|
description |
string |
Description for the output. |
|
jobOutputType |
string:
uri_file |
[Required] Specifies the type of job. |
|
mode | ReadWriteMount |
Output Asset Delivery Mode. |
|
uri |
string |
Output Asset URI. |
UriFolderJobInput
Name | Type | Default value | Description |
---|---|---|---|
description |
string |
Description for the input. |
|
jobInputType |
string:
uri_folder |
[Required] Specifies the type of job. |
|
mode | ReadOnlyMount |
Input Asset Delivery Mode. |
|
uri |
string |
[Required] Input Asset URI. |
UriFolderJobOutput
Name | Type | Default value | Description |
---|---|---|---|
description |
string |
Description for the output. |
|
jobOutputType |
string:
uri_folder |
[Required] Specifies the type of job. |
|
mode | ReadWriteMount |
Output Asset Delivery Mode. |
|
uri |
string |
Output Asset URI. |
UserAssignedIdentity
User assigned identity properties
Name | Type | Description |
---|---|---|
clientId |
string |
The client ID of the assigned identity. |
principalId |
string |
The principal ID of the assigned identity. |
UserIdentity
User identity configuration.
Name | Type | Description |
---|---|---|
identityType | string: |
[Required] Specifies the type of identity framework. |
UseStl
Configure STL Decomposition of the time-series target column.
Name | Type | Description |
---|---|---|
None |
string |
No stl decomposition. |
Season |
string |
|
SeasonTrend |
string |
ValidationMetricType
Metric computation method to use for validation metrics in image tasks.
Name | Type | Description |
---|---|---|
Coco |
string |
Coco metric. |
CocoVoc |
string |
CocoVoc metric. |
None |
string |
No metric. |
Voc |
string |
Voc metric. |
WeekDay
Enum of weekday
Name | Type | Description |
---|---|---|
Friday |
string |
Friday weekday |
Monday |
string |
Monday weekday |
Saturday |
string |
Saturday weekday |
Sunday |
string |
Sunday weekday |
Thursday |
string |
Thursday weekday |
Tuesday |
string |
Tuesday weekday |
Wednesday |
string |
Wednesday weekday |