Azure Resource Manager resource envelope.
Uses Azure REST API version 2025-12-01. In version 2.x of the Azure Native provider, it used API version 2023-04-01.
Other available API versions: 2022-06-01-preview, 2022-10-01, 2022-10-01-preview, 2022-12-01-preview, 2023-02-01-preview, 2023-04-01, 2023-04-01-preview, 2023-06-01-preview, 2023-08-01-preview, 2023-10-01, 2024-01-01-preview, 2024-04-01, 2024-07-01-preview, 2024-10-01, 2024-10-01-preview, 2025-01-01-preview, 2025-04-01, 2025-04-01-preview, 2025-06-01, 2025-07-01-preview, 2025-09-01, 2025-10-01-preview, 2026-01-15-preview, 2026-03-01, 2026-03-15-preview, 2026-05-01, 2026-05-15-preview. These can be accessed by generating a local SDK package using the CLI command pulumi package add azure-native machinelearningservices [ApiVersion]. See the version guide for details.
Example Usage
CreateOrUpdate Schedule.
using System.Collections.Generic;
using System.Linq;
using Pulumi;
using AzureNative = Pulumi.AzureNative;
return await Deployment.RunAsync(() =>
{
var schedule = new AzureNative.MachineLearningServices.Schedule("schedule", new()
{
Name = "string",
Properties = new AzureNative.MachineLearningServices.Inputs.SchedulePropertiesArgs
{
Action = new AzureNative.MachineLearningServices.Inputs.EndpointScheduleActionArgs
{
ActionType = "InvokeBatchEndpoint",
EndpointInvocationDefinition = new Dictionary<string, object?>
{
["9965593e-526f-4b89-bb36-761138cf2794"] = null,
},
},
Description = "string",
DisplayName = "string",
IsEnabled = false,
Properties =
{
{ "string", "string" },
},
Tags =
{
{ "string", "string" },
},
Trigger = new AzureNative.MachineLearningServices.Inputs.CronTriggerArgs
{
EndTime = "string",
Expression = "string",
StartTime = "string",
TimeZone = "string",
TriggerType = "Cron",
},
},
ResourceGroupName = "test-rg",
WorkspaceName = "my-aml-workspace",
});
});
package main
import (
machinelearningservices "github.com/pulumi/pulumi-azure-native-sdk/machinelearningservices/v3"
"github.com/pulumi/pulumi/sdk/v3/go/pulumi"
)
func main() {
pulumi.Run(func(ctx *pulumi.Context) error {
_, err := machinelearningservices.NewSchedule(ctx, "schedule", &machinelearningservices.ScheduleArgs{
Name: pulumi.String("string"),
Properties: &machinelearningservices.SchedulePropertiesArgs{
Action: &machinelearningservices.EndpointScheduleActionArgs{
ActionType: pulumi.String("InvokeBatchEndpoint"),
EndpointInvocationDefinition: pulumi.Any(map[string]interface{}{
"9965593e-526f-4b89-bb36-761138cf2794": nil,
}),
},
Description: pulumi.String("string"),
DisplayName: pulumi.String("string"),
IsEnabled: pulumi.Bool(false),
Properties: pulumi.StringMap{
"string": pulumi.String("string"),
},
Tags: pulumi.StringMap{
"string": pulumi.String("string"),
},
Trigger: &machinelearningservices.CronTriggerArgs{
EndTime: pulumi.String("string"),
Expression: pulumi.String("string"),
StartTime: pulumi.String("string"),
TimeZone: pulumi.String("string"),
TriggerType: pulumi.String("Cron"),
},
},
ResourceGroupName: pulumi.String("test-rg"),
WorkspaceName: pulumi.String("my-aml-workspace"),
})
if err != nil {
return err
}
return nil
})
}
terraform {
required_providers {
azure-native = {
source = "pulumi/azure-native"
}
}
}
resource "azure-native_machinelearningservices_schedule" "schedule" {
lifecycle {
create_before_destroy = true
}
name = "string"
properties = {
action = {
action_type = "InvokeBatchEndpoint"
endpoint_invocation_definition = {
"9965593e-526f-4b89-bb36-761138cf2794" = null
}
}
description = "string"
display_name = "string"
is_enabled = false
properties = {
"string" = "string"
}
tags = {
"string" = "string"
}
trigger = {
end_time = "string"
expression = "string"
start_time = "string"
time_zone = "string"
trigger_type = "Cron"
}
}
resource_group_name = "test-rg"
workspace_name = "my-aml-workspace"
}
package generated_program;
import com.pulumi.Context;
import com.pulumi.Pulumi;
import com.pulumi.core.Output;
import com.pulumi.azurenative.machinelearningservices.Schedule;
import com.pulumi.azurenative.machinelearningservices.ScheduleArgs;
import com.pulumi.azurenative.machinelearningservices.inputs.SchedulePropertiesArgs;
import com.pulumi.azurenative.machinelearningservices.inputs.CreateMonitorActionArgs;
import com.pulumi.azurenative.machinelearningservices.inputs.EndpointScheduleActionArgs;
import com.pulumi.azurenative.machinelearningservices.inputs.JobScheduleActionArgs;
import com.pulumi.azurenative.machinelearningservices.inputs.CronTriggerArgs;
import com.pulumi.azurenative.machinelearningservices.inputs.RecurrenceTriggerArgs;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Map;
import java.io.File;
import java.nio.file.Files;
import java.nio.file.Paths;
public class App {
public static void main(String[] args) {
Pulumi.run(App::stack);
}
public static void stack(Context ctx) {
var schedule = new Schedule("schedule", ScheduleArgs.builder()
.name("string")
.properties(SchedulePropertiesArgs.builder()
.action(EndpointScheduleActionArgs.builder()
.endpointInvocationDefinition(Map.of("9965593e-526f-4b89-bb36-761138cf2794", null))
.build())
.description("string")
.displayName("string")
.isEnabled(false)
.properties(Map.of("string", "string"))
.tags(Map.of("string", "string"))
.trigger(CronTriggerArgs.builder()
.endTime("string")
.expression("string")
.startTime("string")
.timeZone("string")
.build())
.build())
.resourceGroupName("test-rg")
.workspaceName("my-aml-workspace")
.build());
}
}
import * as pulumi from "@pulumi/pulumi";
import * as azure_native from "@pulumi/azure-native";
const schedule = new azure_native.machinelearningservices.Schedule("schedule", {
name: "string",
properties: {
action: {
actionType: "InvokeBatchEndpoint",
endpointInvocationDefinition: {
"9965593e-526f-4b89-bb36-761138cf2794": null,
},
},
description: "string",
displayName: "string",
isEnabled: false,
properties: {
string: "string",
},
tags: {
string: "string",
},
trigger: {
endTime: "string",
expression: "string",
startTime: "string",
timeZone: "string",
triggerType: "Cron",
},
},
resourceGroupName: "test-rg",
workspaceName: "my-aml-workspace",
});
import pulumi
import pulumi_azure_native as azure_native
schedule = azure_native.machinelearningservices.Schedule("schedule",
name="string",
properties={
"action": {
"action_type": "InvokeBatchEndpoint",
"endpoint_invocation_definition": {
"9965593e-526f-4b89-bb36-761138cf2794": None,
},
},
"description": "string",
"display_name": "string",
"is_enabled": False,
"properties": {
"string": "string",
},
"tags": {
"string": "string",
},
"trigger": {
"end_time": "string",
"expression": "string",
"start_time": "string",
"time_zone": "string",
"trigger_type": "Cron",
},
},
resource_group_name="test-rg",
workspace_name="my-aml-workspace")
resources:
schedule:
type: azure-native:machinelearningservices:Schedule
properties:
name: string
properties:
action:
actionType: InvokeBatchEndpoint
endpointInvocationDefinition:
9965593e-526f-4b89-bb36-761138cf2794: null
description: string
displayName: string
isEnabled: false
properties:
string: string
tags:
string: string
trigger:
endTime: string
expression: string
startTime: string
timeZone: string
triggerType: Cron
resourceGroupName: test-rg
workspaceName: my-aml-workspace
Create Schedule Resource
Resources are created with functions called constructors. To learn more about declaring and configuring resources, see Resources.
Constructor syntax
@overload
def Schedule(resource_name: str,
args: ScheduleArgs,
opts: Optional[ResourceOptions] = None)
@overload
def Schedule(resource_name: str,
opts: Optional[ResourceOptions] = None,
properties: Optional[SchedulePropertiesArgs] = None,
resource_group_name: Optional[str] = None,
workspace_name: Optional[str] = None,
name: Optional[str] = None)
public Schedule(String name, ScheduleArgs args)
public Schedule(String name, ScheduleArgs args, CustomResourceOptions options)
type: azure-native:machinelearningservices:Schedule
properties: # The arguments to resource properties.
options: # Bag of options to control resource's behavior.
resource "azure-native_machinelearningservices_schedule" "name" {
# resource properties
}
Parameters
- name
string
- The unique name of the resource.
- args
ScheduleArgs
- The arguments to resource properties.
- opts
CustomResourceOptions
- Bag of options to control resource's behavior.
- resource_name
str
- The unique name of the resource.
- args
ScheduleArgs
- The arguments to resource properties.
- opts
ResourceOptions
- Bag of options to control resource's behavior.
- ctx
Context
- Context object for the current deployment.
- name
string
- The unique name of the resource.
- args
ScheduleArgs
- The arguments to resource properties.
- opts
ResourceOption
- Bag of options to control resource's behavior.
- name
string
- The unique name of the resource.
- args
ScheduleArgs
- The arguments to resource properties.
- opts
CustomResourceOptions
- Bag of options to control resource's behavior.
- name
String
- The unique name of the resource.
- args
ScheduleArgs
- The arguments to resource properties.
- options
CustomResourceOptions
- Bag of options to control resource's behavior.
Constructor example
The following reference example uses placeholder values for all input properties.
var examplescheduleResourceResourceFromMachinelearningservices = new AzureNative.MachineLearningServices.Schedule("examplescheduleResourceResourceFromMachinelearningservices", new()
{
Properties = new AzureNative.MachineLearningServices.Inputs.SchedulePropertiesArgs
{
Action = new AzureNative.MachineLearningServices.Inputs.CreateMonitorActionArgs
{
ActionType = "CreateMonitor",
MonitorDefinition = new AzureNative.MachineLearningServices.Inputs.MonitorDefinitionArgs
{
ComputeConfiguration = new AzureNative.MachineLearningServices.Inputs.MonitorServerlessSparkComputeArgs
{
ComputeIdentity = new AzureNative.MachineLearningServices.Inputs.AmlTokenComputeIdentityArgs
{
ComputeIdentityType = "AmlToken",
},
ComputeType = "ServerlessSpark",
InstanceType = "string",
RuntimeVersion = "string",
},
Signals =
{
{ "string", new AzureNative.MachineLearningServices.Inputs.CustomMonitoringSignalArgs
{
ComponentId = "string",
MetricThresholds = new[]
{
new AzureNative.MachineLearningServices.Inputs.CustomMetricThresholdArgs
{
Metric = "string",
Threshold = new AzureNative.MachineLearningServices.Inputs.MonitoringThresholdArgs
{
Value = 0.0,
},
},
},
SignalType = "Custom",
InputAssets =
{
{ "string", new AzureNative.MachineLearningServices.Inputs.FixedInputDataArgs
{
InputDataType = "Fixed",
JobInputType = "string",
Uri = "string",
Columns =
{
{ "string", "string" },
},
DataContext = "string",
} },
},
Inputs =
{
{ "string", new AzureNative.MachineLearningServices.Inputs.CustomModelJobInputArgs
{
JobInputType = "custom_model",
Uri = "string",
Description = "string",
Mode = "string",
} },
},
NotificationTypes =
{
"string",
},
Properties =
{
{ "string", "string" },
},
} },
},
AlertNotificationSettings = new AzureNative.MachineLearningServices.Inputs.MonitorNotificationSettingsArgs
{
EmailNotificationSettings = new AzureNative.MachineLearningServices.Inputs.MonitorEmailNotificationSettingsArgs
{
Emails = new[]
{
"string",
},
},
},
MonitoringTarget = new AzureNative.MachineLearningServices.Inputs.MonitoringTargetArgs
{
TaskType = "string",
DeploymentId = "string",
ModelId = "string",
},
},
},
Trigger = new AzureNative.MachineLearningServices.Inputs.CronTriggerArgs
{
Expression = "string",
TriggerType = "Cron",
EndTime = "string",
StartTime = "string",
TimeZone = "string",
},
Description = "string",
DisplayName = "string",
IsEnabled = false,
Properties =
{
{ "string", "string" },
},
Tags =
{
{ "string", "string" },
},
},
ResourceGroupName = "string",
WorkspaceName = "string",
Name = "string",
});
example, err := machinelearningservices.NewSchedule(ctx, "examplescheduleResourceResourceFromMachinelearningservices", &machinelearningservices.ScheduleArgs{
Properties: &machinelearningservices.SchedulePropertiesArgs{
Action: &machinelearningservices.CreateMonitorActionArgs{
ActionType: pulumi.String("CreateMonitor"),
MonitorDefinition: &machinelearningservices.MonitorDefinitionArgs{
ComputeConfiguration: &machinelearningservices.MonitorServerlessSparkComputeArgs{
ComputeIdentity: &machinelearningservices.AmlTokenComputeIdentityArgs{
ComputeIdentityType: pulumi.String("AmlToken"),
},
ComputeType: pulumi.String("ServerlessSpark"),
InstanceType: pulumi.String("string"),
RuntimeVersion: pulumi.String("string"),
},
Signals: pulumi.Map{
"string": &machinelearningservices.CustomMonitoringSignalArgs{
ComponentId: pulumi.String("string"),
MetricThresholds: machinelearningservices.CustomMetricThresholdArray{
&machinelearningservices.CustomMetricThresholdArgs{
Metric: pulumi.String("string"),
Threshold: &machinelearningservices.MonitoringThresholdArgs{
Value: pulumi.Float64(0),
},
},
},
SignalType: pulumi.String("Custom"),
InputAssets: pulumi.Map{
"string": &machinelearningservices.FixedInputDataArgs{
InputDataType: pulumi.String("Fixed"),
JobInputType: pulumi.String("string"),
Uri: pulumi.String("string"),
Columns: pulumi.StringMap{
"string": pulumi.String("string"),
},
DataContext: pulumi.String("string"),
},
},
Inputs: pulumi.Map{
"string": &machinelearningservices.CustomModelJobInputArgs{
JobInputType: pulumi.String("custom_model"),
Uri: pulumi.String("string"),
Description: pulumi.String("string"),
Mode: pulumi.String("string"),
},
},
NotificationTypes: pulumi.StringArray{
pulumi.String("string"),
},
Properties: pulumi.StringMap{
"string": pulumi.String("string"),
},
},
},
AlertNotificationSettings: &machinelearningservices.MonitorNotificationSettingsArgs{
EmailNotificationSettings: &machinelearningservices.MonitorEmailNotificationSettingsArgs{
Emails: pulumi.StringArray{
pulumi.String("string"),
},
},
},
MonitoringTarget: &machinelearningservices.MonitoringTargetArgs{
TaskType: pulumi.String("string"),
DeploymentId: pulumi.String("string"),
ModelId: pulumi.String("string"),
},
},
},
Trigger: &machinelearningservices.CronTriggerArgs{
Expression: pulumi.String("string"),
TriggerType: pulumi.String("Cron"),
EndTime: pulumi.String("string"),
StartTime: pulumi.String("string"),
TimeZone: pulumi.String("string"),
},
Description: pulumi.String("string"),
DisplayName: pulumi.String("string"),
IsEnabled: pulumi.Bool(false),
Properties: pulumi.StringMap{
"string": pulumi.String("string"),
},
Tags: pulumi.StringMap{
"string": pulumi.String("string"),
},
},
ResourceGroupName: pulumi.String("string"),
WorkspaceName: pulumi.String("string"),
Name: pulumi.String("string"),
})
resource "azure-native_machinelearningservices_schedule" "examplescheduleResourceResourceFromMachinelearningservices" {
lifecycle {
create_before_destroy = true
}
properties = {
action = {
action_type = "CreateMonitor"
monitor_definition = {
compute_configuration = {
compute_identity = {
compute_identity_type = "AmlToken"
}
compute_type = "ServerlessSpark"
instance_type = "string"
runtime_version = "string"
}
signals = {
"string" = {
component_id = "string"
metric_thresholds = [{
metric = "string"
threshold = {
value = 0
}
}]
signal_type = "Custom"
input_assets = {
"string" = {
input_data_type = "Fixed"
job_input_type = "string"
uri = "string"
columns = {
"string" = "string"
}
data_context = "string"
}
}
inputs = {
"string" = {
job_input_type = "custom_model"
uri = "string"
description = "string"
mode = "string"
}
}
notification_types = ["string"]
properties = {
"string" = "string"
}
}
}
alert_notification_settings = {
email_notification_settings = {
emails = ["string"]
}
}
monitoring_target = {
task_type = "string"
deployment_id = "string"
model_id = "string"
}
}
}
trigger = {
expression = "string"
trigger_type = "Cron"
end_time = "string"
start_time = "string"
time_zone = "string"
}
description = "string"
display_name = "string"
is_enabled = false
properties = {
"string" = "string"
}
tags = {
"string" = "string"
}
}
resource_group_name = "string"
workspace_name = "string"
name = "string"
}
var examplescheduleResourceResourceFromMachinelearningservices = new com.pulumi.azurenative.machinelearningservices.Schedule("examplescheduleResourceResourceFromMachinelearningservices", com.pulumi.azurenative.machinelearningservices.ScheduleArgs.builder()
.properties(com.pulumi.azurenative.machinelearningservices.inputs.SchedulePropertiesArgs.builder()
.action(CreateMonitorActionArgs.builder()
.monitorDefinition(MonitorDefinitionArgs.builder()
.computeConfiguration(MonitorServerlessSparkComputeArgs.builder()
.computeIdentity(AmlTokenComputeIdentityArgs.builder()
.build())
.instanceType("string")
.runtimeVersion("string")
.build())
.signals(Map.of("string", CustomMonitoringSignalArgs.builder()
.componentId("string")
.metricThresholds(CustomMetricThresholdArgs.builder()
.metric("string")
.threshold(MonitoringThresholdArgs.builder()
.value(0.0)
.build())
.build())
.inputAssets(Map.of("string", FixedInputDataArgs.builder()
.jobInputType("string")
.uri("string")
.columns(Map.of("string", "string"))
.dataContext("string")
.build()))
.inputs(Map.of("string", CustomModelJobInputArgs.builder()
.uri("string")
.description("string")
.mode("string")
.build()))
.notificationTypes("string")
.properties(Map.of("string", "string"))
.build()))
.alertNotificationSettings(MonitorNotificationSettingsArgs.builder()
.emailNotificationSettings(MonitorEmailNotificationSettingsArgs.builder()
.emails("string")
.build())
.build())
.monitoringTarget(MonitoringTargetArgs.builder()
.taskType("string")
.deploymentId("string")
.modelId("string")
.build())
.build())
.build())
.trigger(CronTriggerArgs.builder()
.expression("string")
.endTime("string")
.startTime("string")
.timeZone("string")
.build())
.description("string")
.displayName("string")
.isEnabled(false)
.properties(Map.of("string", "string"))
.tags(Map.of("string", "string"))
.build())
.resourceGroupName("string")
.workspaceName("string")
.name("string")
.build());
exampleschedule_resource_resource_from_machinelearningservices = azure_native.machinelearningservices.Schedule("examplescheduleResourceResourceFromMachinelearningservices",
properties={
"action": {
"action_type": "CreateMonitor",
"monitor_definition": {
"compute_configuration": {
"compute_identity": {
"compute_identity_type": "AmlToken",
},
"compute_type": "ServerlessSpark",
"instance_type": "string",
"runtime_version": "string",
},
"signals": {
"string": {
"component_id": "string",
"metric_thresholds": [{
"metric": "string",
"threshold": {
"value": float(0),
},
}],
"signal_type": "Custom",
"input_assets": {
"string": {
"input_data_type": "Fixed",
"job_input_type": "string",
"uri": "string",
"columns": {
"string": "string",
},
"data_context": "string",
},
},
"inputs": {
"string": {
"job_input_type": "custom_model",
"uri": "string",
"description": "string",
"mode": "string",
},
},
"notification_types": ["string"],
"properties": {
"string": "string",
},
},
},
"alert_notification_settings": {
"email_notification_settings": {
"emails": ["string"],
},
},
"monitoring_target": {
"task_type": "string",
"deployment_id": "string",
"model_id": "string",
},
},
},
"trigger": {
"expression": "string",
"trigger_type": "Cron",
"end_time": "string",
"start_time": "string",
"time_zone": "string",
},
"description": "string",
"display_name": "string",
"is_enabled": False,
"properties": {
"string": "string",
},
"tags": {
"string": "string",
},
},
resource_group_name="string",
workspace_name="string",
name="string")
const examplescheduleResourceResourceFromMachinelearningservices = new azure_native.machinelearningservices.Schedule("examplescheduleResourceResourceFromMachinelearningservices", {
properties: {
action: {
actionType: "CreateMonitor",
monitorDefinition: {
computeConfiguration: {
computeIdentity: {
computeIdentityType: "AmlToken",
},
computeType: "ServerlessSpark",
instanceType: "string",
runtimeVersion: "string",
},
signals: {
string: {
componentId: "string",
metricThresholds: [{
metric: "string",
threshold: {
value: 0,
},
}],
signalType: "Custom",
inputAssets: {
string: {
inputDataType: "Fixed",
jobInputType: "string",
uri: "string",
columns: {
string: "string",
},
dataContext: "string",
},
},
inputs: {
string: {
jobInputType: "custom_model",
uri: "string",
description: "string",
mode: "string",
},
},
notificationTypes: ["string"],
properties: {
string: "string",
},
},
},
alertNotificationSettings: {
emailNotificationSettings: {
emails: ["string"],
},
},
monitoringTarget: {
taskType: "string",
deploymentId: "string",
modelId: "string",
},
},
},
trigger: {
expression: "string",
triggerType: "Cron",
endTime: "string",
startTime: "string",
timeZone: "string",
},
description: "string",
displayName: "string",
isEnabled: false,
properties: {
string: "string",
},
tags: {
string: "string",
},
},
resourceGroupName: "string",
workspaceName: "string",
name: "string",
});
type: azure-native:machinelearningservices:Schedule
properties:
name: string
properties:
action:
actionType: CreateMonitor
monitorDefinition:
alertNotificationSettings:
emailNotificationSettings:
emails:
- string
computeConfiguration:
computeIdentity:
computeIdentityType: AmlToken
computeType: ServerlessSpark
instanceType: string
runtimeVersion: string
monitoringTarget:
deploymentId: string
modelId: string
taskType: string
signals:
string:
componentId: string
inputAssets:
string:
columns:
string: string
dataContext: string
inputDataType: Fixed
jobInputType: string
uri: string
inputs:
string:
description: string
jobInputType: custom_model
mode: string
uri: string
metricThresholds:
- metric: string
threshold:
value: 0
notificationTypes:
- string
properties:
string: string
signalType: Custom
description: string
displayName: string
isEnabled: false
properties:
string: string
tags:
string: string
trigger:
endTime: string
expression: string
startTime: string
timeZone: string
triggerType: Cron
resourceGroupName: string
workspaceName: string
Schedule Resource Properties
To learn more about resource properties and how to use them, see Inputs and Outputs in the Architecture and Concepts docs.
In Python, inputs that are objects can be passed either as argument classes or as dictionary literals.
The Schedule resource accepts the following input properties:
Outputs
All input properties are implicitly available as output properties. Additionally, the Schedule resource produces the following output properties:
- AzureApiVersion
string
- The Azure API version of the resource.
- Id
string
- The provider-assigned unique ID for this managed resource.
- SystemData
SystemDataResponse
- 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"
- azure_api_version
string
- The Azure API version of the resource.
- id
string
- The provider-assigned unique ID for this managed resource.
- system_data
object
- 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"
- azureApiVersion
String
- The Azure API version of the resource.
- id
String
- The provider-assigned unique ID for this managed resource.
- systemData
SystemDataResponse
- 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"
- azureApiVersion
string
- The Azure API version of the resource.
- id
string
- The provider-assigned unique ID for this managed resource.
- systemData
SystemDataResponse
- 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"
- azure_api_version
str
- The Azure API version of the resource.
- id
str
- The provider-assigned unique ID for this managed resource.
- system_data
SystemDataResponse
- Azure Resource Manager metadata containing createdBy and modifiedBy information.
- type
str
- The type of the resource. E.g. "Microsoft.Compute/virtualMachines" or "Microsoft.Storage/storageAccounts"
- azureApiVersion
String
- The Azure API version of the resource.
- id
String
- The provider-assigned unique ID for this managed resource.
- systemData
Property Map
- 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"
Look up Existing Schedule Resource
Get an existing Schedule resource’s state with the given name, ID, and optional extra properties used to qualify the lookup.
public static Schedule get(String name, Output<String> id, CustomResourceOptions options)
resources: existingSchedule: type: azure-native:machinelearningservices:Schedule get: id: ${id}
import {
to = azure-native_machinelearningservices_schedule.example
id = "${id}"
}
- name
string
- The unique name of the resulting resource.
- id
Input<ID>
- The unique provider ID of the resource to look up.
- opts
CustomResourceOptions
- Bag of options to control resource's behavior.
- resource_name
str
- The unique name of the resulting resource.
- id
str
- The unique provider ID of the resource to look up.
- opts
ResourceOptions
- Bag of options to control resource's behavior.
- ctx
Context
- Context object for the current deployment.
- name
string
- The unique name of the resulting resource.
- id
IDInput
- The unique provider ID of the resource to look up.
- state
ScheduleState
- Pass nil. This resource has no state arguments.
- opts
ResourceOption
- Bag of options to control resource's behavior.
- name
string
- The unique name of the resulting resource.
- id
Input<string>
- The unique provider ID of the resource to look up.
- opts
CustomResourceOptions
- Bag of options to control resource's behavior.
- name
String
- The unique name of the resulting resource.
- id
Output<String>
- The unique provider ID of the resource to look up.
- options
CustomResourceOptions
- Bag of options to control resource's behavior.
Supporting Types
AllFeatures, AllFeaturesArgs
AllFeaturesResponse
AllNodes, AllNodesArgs
All nodes means the service will be running on all of the nodes of the job
AllNodesResponse
All nodes means the service will be running on all of the nodes of the job
AmlToken, AmlTokenArgs
AML Token identity configuration.
AmlTokenComputeIdentity, AmlTokenComputeIdentityArgs
AML token compute identity definition.
AmlTokenComputeIdentityResponse
AML token compute identity definition.
AmlTokenResponse
AML Token identity configuration.
AutoForecastHorizon, AutoForecastHorizonArgs
Forecast horizon determined automatically by system.
AutoForecastHorizonResponse
Forecast horizon determined automatically by system.
AutoMLJob, AutoMLJobArgs
AutoMLJob class.
Use this class for executing AutoML tasks like Classification/Regression etc.
See TaskType enum for all the tasks supported.
- TaskDetails
Pulumi.AzureNative.MachineLearningServices.Inputs.Classification | Pulumi.AzureNative.MachineLearningServices.Inputs.Forecasting | Pulumi.AzureNative.MachineLearningServices.Inputs.ImageClassification | Pulumi.AzureNative.MachineLearningServices.Inputs.ImageClassificationMultilabel | Pulumi.AzureNative.MachineLearningServices.Inputs.ImageInstanceSegmentation | Pulumi.AzureNative.MachineLearningServices.Inputs.ImageObjectDetection | Pulumi.AzureNative.MachineLearningServices.Inputs.Regression | Pulumi.AzureNative.MachineLearningServices.Inputs.TextClassification | Pulumi.AzureNative.MachineLearningServices.Inputs.TextClassificationMultilabel | Pulumi.AzureNative.MachineLearningServices.Inputs.TextNer
- [Required] This represents scenario which can be one of Tables/NLP/Image
- 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
Dictionary<string, string>
- Environment variables included in the job.
- ExperimentName
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- Identity
Pulumi.AzureNative.MachineLearningServices.Inputs.AmlToken | Pulumi.AzureNative.MachineLearningServices.Inputs.ManagedIdentity | Pulumi.AzureNative.MachineLearningServices.Inputs.UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- IsArchived
bool
- Is the asset archived?
- NotificationSetting
Pulumi.AzureNative.MachineLearningServices.Inputs.NotificationSetting
- Notification setting for the job
- Outputs
Dictionary<string, object>
- Mapping of output data bindings used in the job.
- Properties
Dictionary<string, string>
- The asset property dictionary.
- QueueSettings
Pulumi.AzureNative.MachineLearningServices.Inputs.QueueSettings
- Queue settings for the job
- Resources
Pulumi.AzureNative.MachineLearningServices.Inputs.JobResourceConfiguration
- Compute Resource configuration for the job.
- Services
Dictionary<string, Pulumi.AzureNative.MachineLearningServices.Inputs.JobService>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- Tags
Dictionary<string, string>
- Tag dictionary. Tags can be added, removed, and updated.
- task_details
object | object | object | object | object | object | object | object | object | object
- [Required] This represents scenario which can be one of Tables/NLP/Image
- component_id
string
- ARM resource ID of the component resource.
- compute_id
string
- ARM resource ID of the compute resource.
- description
string
- The asset description text.
- display_name
string
- Display name of job.
- environment_id
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.
- environment_variables
map(string)
- Environment variables included in the job.
- experiment_name
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
object | object | object
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- is_archived
bool
- Is the asset archived?
- notification_setting
object
- Notification setting for the job
- outputs
map(object | object | object | object | object | object)
- Mapping of output data bindings used in the job.
- properties
map(string)
- The asset property dictionary.
- queue_settings
object
- Queue settings for the job
- resources
object
- Compute Resource configuration for the job.
- services
map(object)
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
map(string)
- Tag dictionary. Tags can be added, removed, and updated.
- taskDetails
Classification | Forecasting | ImageClassification | ImageClassificationMultilabel | ImageInstanceSegmentation | ImageObjectDetection | Regression | TextClassification | TextClassificationMultilabel | TextNer
- [Required] This represents scenario which can be one of Tables/NLP/Image
- 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
Map<String,String>
- Environment variables included in the job.
- experimentName
String
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlToken | ManagedIdentity | UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- isArchived
Boolean
- Is the asset archived?
- notificationSetting
NotificationSetting
- Notification setting for the job
- outputs
Map<String,Object>
- Mapping of output data bindings used in the job.
- properties
Map<String,String>
- The asset property dictionary.
- queueSettings
QueueSettings
- Queue settings for the job
- resources
JobResourceConfiguration
- Compute Resource configuration for the job.
- services
Map<String,JobService>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Map<String,String>
- Tag dictionary. Tags can be added, removed, and updated.
- task_details
Classification | Forecasting | ImageClassification | ImageClassificationMultilabel | ImageInstanceSegmentation | ImageObjectDetection | Regression | TextClassification | TextClassificationMultilabel | TextNer
- [Required] This represents scenario which can be one of Tables/NLP/Image
- component_id
str
- ARM resource ID of the component resource.
- compute_id
str
- ARM resource ID of the compute resource.
- description
str
- The asset description text.
- display_name
str
- Display name of job.
- environment_id
str
- 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.
- environment_variables
Mapping[str, str]
- Environment variables included in the job.
- experiment_name
str
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlToken | ManagedIdentity | UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- is_archived
bool
- Is the asset archived?
- notification_setting
NotificationSetting
- Notification setting for the job
- outputs
Mapping[str, Union[CustomModelJobOutput, MLFlowModelJobOutput, MLTableJobOutput, TritonModelJobOutput, UriFileJobOutput, UriFolderJobOutput]]
- Mapping of output data bindings used in the job.
- properties
Mapping[str, str]
- The asset property dictionary.
- queue_settings
QueueSettings
- Queue settings for the job
- resources
JobResourceConfiguration
- Compute Resource configuration for the job.
- services
Mapping[str, JobService]
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Mapping[str, str]
- Tag dictionary. Tags can be added, removed, and updated.
AutoMLJobResponse
AutoMLJob class.
Use this class for executing AutoML tasks like Classification/Regression etc.
See TaskType enum for all the tasks supported.
- Status
string
- Status of the job.
- TaskDetails
Pulumi.AzureNative.MachineLearningServices.Inputs.ClassificationResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.ForecastingResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.ImageClassificationResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.ImageClassificationMultilabelResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.ImageInstanceSegmentationResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.ImageObjectDetectionResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.RegressionResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.TextClassificationResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.TextClassificationMultilabelResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.TextNerResponse
- [Required] This represents scenario which can be one of Tables/NLP/Image
- 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
Dictionary<string, string>
- Environment variables included in the job.
- ExperimentName
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- Identity
Pulumi.AzureNative.MachineLearningServices.Inputs.AmlTokenResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.ManagedIdentityResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- IsArchived
bool
- Is the asset archived?
- NotificationSetting
Pulumi.AzureNative.MachineLearningServices.Inputs.NotificationSettingResponse
- Notification setting for the job
- Outputs
Dictionary<string, object>
- Mapping of output data bindings used in the job.
- Properties
Dictionary<string, string>
- The asset property dictionary.
- QueueSettings
Pulumi.AzureNative.MachineLearningServices.Inputs.QueueSettingsResponse
- Queue settings for the job
- Resources
Pulumi.AzureNative.MachineLearningServices.Inputs.JobResourceConfigurationResponse
- Compute Resource configuration for the job.
- Services
Dictionary<string, Pulumi.AzureNative.MachineLearningServices.Inputs.JobServiceResponse>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- Tags
Dictionary<string, string>
- Tag dictionary. Tags can be added, removed, and updated.
- status
string
- Status of the job.
- task_details
object | object | object | object | object | object | object | object | object | object
- [Required] This represents scenario which can be one of Tables/NLP/Image
- component_id
string
- ARM resource ID of the component resource.
- compute_id
string
- ARM resource ID of the compute resource.
- description
string
- The asset description text.
- display_name
string
- Display name of job.
- environment_id
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.
- environment_variables
map(string)
- Environment variables included in the job.
- experiment_name
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
object | object | object
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- is_archived
bool
- Is the asset archived?
- notification_setting
object
- Notification setting for the job
- outputs
map(object | object | object | object | object | object)
- Mapping of output data bindings used in the job.
- properties
map(string)
- The asset property dictionary.
- queue_settings
object
- Queue settings for the job
- resources
object
- Compute Resource configuration for the job.
- services
map(object)
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
map(string)
- Tag dictionary. Tags can be added, removed, and updated.
- status
String
- Status of the job.
- taskDetails
ClassificationResponse | ForecastingResponse | ImageClassificationResponse | ImageClassificationMultilabelResponse | ImageInstanceSegmentationResponse | ImageObjectDetectionResponse | RegressionResponse | TextClassificationResponse | TextClassificationMultilabelResponse | TextNerResponse
- [Required] This represents scenario which can be one of Tables/NLP/Image
- 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
Map<String,String>
- Environment variables included in the job.
- experimentName
String
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlTokenResponse | ManagedIdentityResponse | UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- isArchived
Boolean
- Is the asset archived?
- notificationSetting
NotificationSettingResponse
- Notification setting for the job
- outputs
Map<String,Object>
- Mapping of output data bindings used in the job.
- properties
Map<String,String>
- The asset property dictionary.
- queueSettings
QueueSettingsResponse
- Queue settings for the job
- resources
JobResourceConfigurationResponse
- Compute Resource configuration for the job.
- services
Map<String,JobServiceResponse>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Map<String,String>
- Tag dictionary. Tags can be added, removed, and updated.
- status
str
- Status of the job.
- task_details
ClassificationResponse | ForecastingResponse | ImageClassificationResponse | ImageClassificationMultilabelResponse | ImageInstanceSegmentationResponse | ImageObjectDetectionResponse | RegressionResponse | TextClassificationResponse | TextClassificationMultilabelResponse | TextNerResponse
- [Required] This represents scenario which can be one of Tables/NLP/Image
- component_id
str
- ARM resource ID of the component resource.
- compute_id
str
- ARM resource ID of the compute resource.
- description
str
- The asset description text.
- display_name
str
- Display name of job.
- environment_id
str
- 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.
- environment_variables
Mapping[str, str]
- Environment variables included in the job.
- experiment_name
str
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlTokenResponse | ManagedIdentityResponse | UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- is_archived
bool
- Is the asset archived?
- notification_setting
NotificationSettingResponse
- Notification setting for the job
- outputs
Mapping[str, Union[CustomModelJobOutputResponse, MLFlowModelJobOutputResponse, MLTableJobOutputResponse, TritonModelJobOutputResponse, UriFileJobOutputResponse, UriFolderJobOutputResponse]]
- Mapping of output data bindings used in the job.
- properties
Mapping[str, str]
- The asset property dictionary.
- queue_settings
QueueSettingsResponse
- Queue settings for the job
- resources
JobResourceConfigurationResponse
- Compute Resource configuration for the job.
- services
Mapping[str, JobServiceResponse]
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Mapping[str, str]
- Tag dictionary. Tags can be added, removed, and updated.
AutoNCrossValidations, AutoNCrossValidationsArgs
N-Cross validations determined automatically.
AutoNCrossValidationsResponse
N-Cross validations determined automatically.
AutoSeasonality, AutoSeasonalityArgs
AutoSeasonalityResponse
AutoTargetLags, AutoTargetLagsArgs
AutoTargetLagsResponse
AutoTargetRollingWindowSize, AutoTargetRollingWindowSizeArgs
Target lags rolling window determined automatically.
AutoTargetRollingWindowSizeResponse
Target lags rolling window determined automatically.
AzureDevOpsWebhook, AzureDevOpsWebhookArgs
Webhook details specific for Azure DevOps
- EventType
string
- Send callback on a specified notification event
- EventType
string
- Send callback on a specified notification event
- event_type
string
- Send callback on a specified notification event
- eventType
String
- Send callback on a specified notification event
- eventType
string
- Send callback on a specified notification event
- event_type
str
- Send callback on a specified notification event
- eventType
String
- Send callback on a specified notification event
AzureDevOpsWebhookResponse
Webhook details specific for Azure DevOps
- EventType
string
- Send callback on a specified notification event
- EventType
string
- Send callback on a specified notification event
- event_type
string
- Send callback on a specified notification event
- eventType
String
- Send callback on a specified notification event
- eventType
string
- Send callback on a specified notification event
- event_type
str
- Send callback on a specified notification event
- eventType
String
- Send callback on a specified notification event
BanditPolicy, BanditPolicyArgs
Defines an early termination policy based on slack criteria, and a frequency and delay interval for evaluation
- DelayEvaluation
int
- Number of intervals by which to delay the first evaluation.
- EvaluationInterval
int
- Interval (number of runs) between policy evaluations.
- SlackAmount
double
- Absolute distance allowed from the best performing run.
- SlackFactor
double
- Ratio of the allowed distance from the best performing run.
- DelayEvaluation
int
- Number of intervals by which to delay the first evaluation.
- EvaluationInterval
int
- Interval (number of runs) between policy evaluations.
- SlackAmount
float64
- Absolute distance allowed from the best performing run.
- SlackFactor
float64
- Ratio of the allowed distance from the best performing run.
- delay_evaluation
number
- Number of intervals by which to delay the first evaluation.
- evaluation_interval
number
- Interval (number of runs) between policy evaluations.
- slack_amount
number
- Absolute distance allowed from the best performing run.
- slack_factor
number
- Ratio of the allowed distance from the best performing run.
- delayEvaluation
Integer
- Number of intervals by which to delay the first evaluation.
- evaluationInterval
Integer
- Interval (number of runs) between policy evaluations.
- slackAmount
Double
- Absolute distance allowed from the best performing run.
- slackFactor
Double
- Ratio of the allowed distance from the best performing run.
- delayEvaluation
number
- Number of intervals by which to delay the first evaluation.
- evaluationInterval
number
- Interval (number of runs) between policy evaluations.
- slackAmount
number
- Absolute distance allowed from the best performing run.
- slackFactor
number
- Ratio of the allowed distance from the best performing run.
- delay_evaluation
int
- Number of intervals by which to delay the first evaluation.
- evaluation_interval
int
- Interval (number of runs) between policy evaluations.
- slack_amount
float
- Absolute distance allowed from the best performing run.
- slack_factor
float
- Ratio of the allowed distance from the best performing run.
- delayEvaluation
Number
- Number of intervals by which to delay the first evaluation.
- evaluationInterval
Number
- Interval (number of runs) between policy evaluations.
- slackAmount
Number
- Absolute distance allowed from the best performing run.
- slackFactor
Number
- Ratio of the allowed distance from the best performing run.
BanditPolicyResponse
Defines an early termination policy based on slack criteria, and a frequency and delay interval for evaluation
- DelayEvaluation
int
- Number of intervals by which to delay the first evaluation.
- EvaluationInterval
int
- Interval (number of runs) between policy evaluations.
- SlackAmount
double
- Absolute distance allowed from the best performing run.
- SlackFactor
double
- Ratio of the allowed distance from the best performing run.
- DelayEvaluation
int
- Number of intervals by which to delay the first evaluation.
- EvaluationInterval
int
- Interval (number of runs) between policy evaluations.
- SlackAmount
float64
- Absolute distance allowed from the best performing run.
- SlackFactor
float64
- Ratio of the allowed distance from the best performing run.
- delay_evaluation
number
- Number of intervals by which to delay the first evaluation.
- evaluation_interval
number
- Interval (number of runs) between policy evaluations.
- slack_amount
number
- Absolute distance allowed from the best performing run.
- slack_factor
number
- Ratio of the allowed distance from the best performing run.
- delayEvaluation
Integer
- Number of intervals by which to delay the first evaluation.
- evaluationInterval
Integer
- Interval (number of runs) between policy evaluations.
- slackAmount
Double
- Absolute distance allowed from the best performing run.
- slackFactor
Double
- Ratio of the allowed distance from the best performing run.
- delayEvaluation
number
- Number of intervals by which to delay the first evaluation.
- evaluationInterval
number
- Interval (number of runs) between policy evaluations.
- slackAmount
number
- Absolute distance allowed from the best performing run.
- slackFactor
number
- Ratio of the allowed distance from the best performing run.
- delay_evaluation
int
- Number of intervals by which to delay the first evaluation.
- evaluation_interval
int
- Interval (number of runs) between policy evaluations.
- slack_amount
float
- Absolute distance allowed from the best performing run.
- slack_factor
float
- Ratio of the allowed distance from the best performing run.
- delayEvaluation
Number
- Number of intervals by which to delay the first evaluation.
- evaluationInterval
Number
- Interval (number of runs) between policy evaluations.
- slackAmount
Number
- Absolute distance allowed from the best performing run.
- slackFactor
Number
- Ratio of the allowed distance from the best performing run.
BayesianSamplingAlgorithm, BayesianSamplingAlgorithmArgs
Defines a Sampling Algorithm that generates values based on previous values
BayesianSamplingAlgorithmResponse
Defines a Sampling Algorithm that generates values based on previous values
Enum for all classification models supported by AutoML.
| Name | Value | Description |
|---|
| TextTargetEncoder | TextTargetEncoder | Target encoding for text data. |
| OneHotEncoder | OneHotEncoder | Ohe hot encoding creates a binary feature transformation. |
| CatTargetEncoder | CatTargetEncoder | Target encoding for categorical data. |
| TfIdf | TfIdf | Tf-Idf stands for, term-frequency times inverse document-frequency. This is a common term weighting scheme for identifying information from documents. |
| WoETargetEncoder | WoETargetEncoder | 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. |
| LabelEncoder | LabelEncoder | Label encoder converts labels/categorical variables in a numerical form. |
| WordEmbedding | WordEmbedding | Word embedding helps represents words or phrases as a vector, or a series of numbers. |
| NaiveBayes | NaiveBayes | Naive Bayes is a classified that is used for classification of discrete features that are categorically distributed. |
| CountVectorizer | CountVectorizer | Count Vectorizer converts a collection of text documents to a matrix of token counts. |
| HashOneHotEncoder | HashOneHotEncoder | Hashing One Hot Encoder can turn categorical variables into a limited number of new features. This is often used for high-cardinality categorical features. |
| Name | Value | Description |
|---|
| BlockedTransformersTextTargetEncoder | TextTargetEncoder | Target encoding for text data. |
| BlockedTransformersOneHotEncoder | OneHotEncoder | Ohe hot encoding creates a binary feature transformation. |
| BlockedTransformersCatTargetEncoder | CatTargetEncoder | Target encoding for categorical data. |
| BlockedTransformersTfIdf | TfIdf | Tf-Idf stands for, term-frequency times inverse document-frequency. This is a common term weighting scheme for identifying information from documents. |
| BlockedTransformersWoETargetEncoder | WoETargetEncoder | 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. |
| BlockedTransformersLabelEncoder | LabelEncoder | Label encoder converts labels/categorical variables in a numerical form. |
| BlockedTransformersWordEmbedding | WordEmbedding | Word embedding helps represents words or phrases as a vector, or a series of numbers. |
| BlockedTransformersNaiveBayes | NaiveBayes | Naive Bayes is a classified that is used for classification of discrete features that are categorically distributed. |
| BlockedTransformersCountVectorizer | CountVectorizer | Count Vectorizer converts a collection of text documents to a matrix of token counts. |
| BlockedTransformersHashOneHotEncoder | HashOneHotEncoder | Hashing One Hot Encoder can turn categorical variables into a limited number of new features. This is often used for high-cardinality categorical features. |
| Value | Description |
|---|
TextTargetEncoder | Target encoding for text data. |
OneHotEncoder | Ohe hot encoding creates a binary feature transformation. |
CatTargetEncoder | Target encoding for categorical data. |
TfIdf | Tf-Idf stands for, term-frequency times inverse document-frequency. This is a common term weighting scheme for identifying information from documents. |
WoETargetEncoder | 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. |
LabelEncoder | Label encoder converts labels/categorical variables in a numerical form. |
WordEmbedding | Word embedding helps represents words or phrases as a vector, or a series of numbers. |
NaiveBayes | Naive Bayes is a classified that is used for classification of discrete features that are categorically distributed. |
CountVectorizer | Count Vectorizer converts a collection of text documents to a matrix of token counts. |
HashOneHotEncoder | Hashing One Hot Encoder can turn categorical variables into a limited number of new features. This is often used for high-cardinality categorical features. |
| Name | Value | Description |
|---|
| TextTargetEncoder | TextTargetEncoder | Target encoding for text data. |
| OneHotEncoder | OneHotEncoder | Ohe hot encoding creates a binary feature transformation. |
| CatTargetEncoder | CatTargetEncoder | Target encoding for categorical data. |
| TfIdf | TfIdf | Tf-Idf stands for, term-frequency times inverse document-frequency. This is a common term weighting scheme for identifying information from documents. |
| WoETargetEncoder | WoETargetEncoder | 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. |
| LabelEncoder | LabelEncoder | Label encoder converts labels/categorical variables in a numerical form. |
| WordEmbedding | WordEmbedding | Word embedding helps represents words or phrases as a vector, or a series of numbers. |
| NaiveBayes | NaiveBayes | Naive Bayes is a classified that is used for classification of discrete features that are categorically distributed. |
| CountVectorizer | CountVectorizer | Count Vectorizer converts a collection of text documents to a matrix of token counts. |
| HashOneHotEncoder | HashOneHotEncoder | Hashing One Hot Encoder can turn categorical variables into a limited number of new features. This is often used for high-cardinality categorical features. |
| Name | Value | Description |
|---|
| TextTargetEncoder | TextTargetEncoder | Target encoding for text data. |
| OneHotEncoder | OneHotEncoder | Ohe hot encoding creates a binary feature transformation. |
| CatTargetEncoder | CatTargetEncoder | Target encoding for categorical data. |
| TfIdf | TfIdf | Tf-Idf stands for, term-frequency times inverse document-frequency. This is a common term weighting scheme for identifying information from documents. |
| WoETargetEncoder | WoETargetEncoder | 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. |
| LabelEncoder | LabelEncoder | Label encoder converts labels/categorical variables in a numerical form. |
| WordEmbedding | WordEmbedding | Word embedding helps represents words or phrases as a vector, or a series of numbers. |
| NaiveBayes | NaiveBayes | Naive Bayes is a classified that is used for classification of discrete features that are categorically distributed. |
| CountVectorizer | CountVectorizer | Count Vectorizer converts a collection of text documents to a matrix of token counts. |
| HashOneHotEncoder | HashOneHotEncoder | Hashing One Hot Encoder can turn categorical variables into a limited number of new features. This is often used for high-cardinality categorical features. |
| Name | Value | Description |
|---|
| TEXT_TARGET_ENCODER | TextTargetEncoder | Target encoding for text data. |
| ONE_HOT_ENCODER | OneHotEncoder | Ohe hot encoding creates a binary feature transformation. |
| CAT_TARGET_ENCODER | CatTargetEncoder | Target encoding for categorical data. |
| TF_IDF | TfIdf | Tf-Idf stands for, term-frequency times inverse document-frequency. This is a common term weighting scheme for identifying information from documents. |
| WO_E_TARGET_ENCODER | WoETargetEncoder | 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. |
| LABEL_ENCODER | LabelEncoder | Label encoder converts labels/categorical variables in a numerical form. |
| WORD_EMBEDDING | WordEmbedding | Word embedding helps represents words or phrases as a vector, or a series of numbers. |
| NAIVE_BAYES | NaiveBayes | Naive Bayes is a classified that is used for classification of discrete features that are categorically distributed. |
| COUNT_VECTORIZER | CountVectorizer | Count Vectorizer converts a collection of text documents to a matrix of token counts. |
| HASH_ONE_HOT_ENCODER | HashOneHotEncoder | Hashing One Hot Encoder can turn categorical variables into a limited number of new features. This is often used for high-cardinality categorical features. |
| Value | Description |
|---|
TextTargetEncoder | Target encoding for text data. |
OneHotEncoder | Ohe hot encoding creates a binary feature transformation. |
CatTargetEncoder | Target encoding for categorical data. |
TfIdf | Tf-Idf stands for, term-frequency times inverse document-frequency. This is a common term weighting scheme for identifying information from documents. |
WoETargetEncoder | 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. |
LabelEncoder | Label encoder converts labels/categorical variables in a numerical form. |
WordEmbedding | Word embedding helps represents words or phrases as a vector, or a series of numbers. |
NaiveBayes | Naive Bayes is a classified that is used for classification of discrete features that are categorically distributed. |
CountVectorizer | Count Vectorizer converts a collection of text documents to a matrix of token counts. |
HashOneHotEncoder | Hashing One Hot Encoder can turn categorical variables into a limited number of new features. This is often used for high-cardinality categorical features. |
CategoricalDataDriftMetric
[Required] The categorical data drift metric to calculate.
| Name | Value | Description |
|---|
| JensenShannonDistance | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| PopulationStabilityIndex | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| PearsonsChiSquaredTest | PearsonsChiSquaredTest | The Pearsons Chi Squared Test metric. |
| Name | Value | Description |
|---|
| CategoricalDataDriftMetricJensenShannonDistance | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| CategoricalDataDriftMetricPopulationStabilityIndex | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| CategoricalDataDriftMetricPearsonsChiSquaredTest | PearsonsChiSquaredTest | The Pearsons Chi Squared Test metric. |
| Value | Description |
|---|
JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
PearsonsChiSquaredTest | The Pearsons Chi Squared Test metric. |
| Name | Value | Description |
|---|
| JensenShannonDistance | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| PopulationStabilityIndex | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| PearsonsChiSquaredTest | PearsonsChiSquaredTest | The Pearsons Chi Squared Test metric. |
| Name | Value | Description |
|---|
| JensenShannonDistance | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| PopulationStabilityIndex | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| PearsonsChiSquaredTest | PearsonsChiSquaredTest | The Pearsons Chi Squared Test metric. |
| Name | Value | Description |
|---|
| JENSEN_SHANNON_DISTANCE | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| POPULATION_STABILITY_INDEX | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| PEARSONS_CHI_SQUARED_TEST | PearsonsChiSquaredTest | The Pearsons Chi Squared Test metric. |
| Value | Description |
|---|
JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
PearsonsChiSquaredTest | The Pearsons Chi Squared Test metric. |
CategoricalDataDriftMetricThreshold, CategoricalDataDriftMetricThresholdArgs
CategoricalDataDriftMetricThresholdResponse
- metric
string
- [Required] The categorical data drift metric to calculate.
- threshold
object
- The threshold value. If null, a default value will be set depending on the selected metric.
- metric
String
- [Required] The categorical data drift metric to calculate.
- threshold
Property Map
- The threshold value. If null, a default value will be set depending on the selected metric.
CategoricalDataQualityMetric
[Required] The categorical data quality metric to calculate.
| Name | Value | Description |
|---|
| NullValueRate | NullValueRate | Calculates the rate of null values. |
| DataTypeErrorRate | DataTypeErrorRate | Calculates the rate of data type errors. |
| OutOfBoundsRate | OutOfBoundsRate | Calculates the rate values are out of bounds. |
| Name | Value | Description |
|---|
| CategoricalDataQualityMetricNullValueRate | NullValueRate | Calculates the rate of null values. |
| CategoricalDataQualityMetricDataTypeErrorRate | DataTypeErrorRate | Calculates the rate of data type errors. |
| CategoricalDataQualityMetricOutOfBoundsRate | OutOfBoundsRate | Calculates the rate values are out of bounds. |
| Value | Description |
|---|
NullValueRate | Calculates the rate of null values. |
DataTypeErrorRate | Calculates the rate of data type errors. |
OutOfBoundsRate | Calculates the rate values are out of bounds. |
| Name | Value | Description |
|---|
| NullValueRate | NullValueRate | Calculates the rate of null values. |
| DataTypeErrorRate | DataTypeErrorRate | Calculates the rate of data type errors. |
| OutOfBoundsRate | OutOfBoundsRate | Calculates the rate values are out of bounds. |
| Name | Value | Description |
|---|
| NullValueRate | NullValueRate | Calculates the rate of null values. |
| DataTypeErrorRate | DataTypeErrorRate | Calculates the rate of data type errors. |
| OutOfBoundsRate | OutOfBoundsRate | Calculates the rate values are out of bounds. |
| Name | Value | Description |
|---|
| NULL_VALUE_RATE | NullValueRate | Calculates the rate of null values. |
| DATA_TYPE_ERROR_RATE | DataTypeErrorRate | Calculates the rate of data type errors. |
| OUT_OF_BOUNDS_RATE | OutOfBoundsRate | Calculates the rate values are out of bounds. |
| Value | Description |
|---|
NullValueRate | Calculates the rate of null values. |
DataTypeErrorRate | Calculates the rate of data type errors. |
OutOfBoundsRate | Calculates the rate values are out of bounds. |
CategoricalDataQualityMetricThreshold, CategoricalDataQualityMetricThresholdArgs
CategoricalDataQualityMetricThresholdResponse
- metric
string
- [Required] The categorical data quality metric to calculate.
- threshold
object
- The threshold value. If null, a default value will be set depending on the selected metric.
- metric
String
- [Required] The categorical data quality metric to calculate.
- threshold
Property Map
- The threshold value. If null, a default value will be set depending on the selected metric.
CategoricalPredictionDriftMetric
[Required] The categorical prediction drift metric to calculate.
| Name | Value | Description |
|---|
| JensenShannonDistance | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| PopulationStabilityIndex | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| PearsonsChiSquaredTest | PearsonsChiSquaredTest | The Pearsons Chi Squared Test metric. |
| Name | Value | Description |
|---|
| CategoricalPredictionDriftMetricJensenShannonDistance | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| CategoricalPredictionDriftMetricPopulationStabilityIndex | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| CategoricalPredictionDriftMetricPearsonsChiSquaredTest | PearsonsChiSquaredTest | The Pearsons Chi Squared Test metric. |
| Value | Description |
|---|
JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
PearsonsChiSquaredTest | The Pearsons Chi Squared Test metric. |
| Name | Value | Description |
|---|
| JensenShannonDistance | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| PopulationStabilityIndex | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| PearsonsChiSquaredTest | PearsonsChiSquaredTest | The Pearsons Chi Squared Test metric. |
| Name | Value | Description |
|---|
| JensenShannonDistance | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| PopulationStabilityIndex | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| PearsonsChiSquaredTest | PearsonsChiSquaredTest | The Pearsons Chi Squared Test metric. |
| Name | Value | Description |
|---|
| JENSEN_SHANNON_DISTANCE | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| POPULATION_STABILITY_INDEX | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| PEARSONS_CHI_SQUARED_TEST | PearsonsChiSquaredTest | The Pearsons Chi Squared Test metric. |
| Value | Description |
|---|
JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
PearsonsChiSquaredTest | The Pearsons Chi Squared Test metric. |
CategoricalPredictionDriftMetricThreshold, CategoricalPredictionDriftMetricThresholdArgs
CategoricalPredictionDriftMetricThresholdResponse
- metric
string
- [Required] The categorical prediction drift metric to calculate.
- threshold
object
- The threshold value. If null, a default value will be set depending on the selected metric.
- metric
String
- [Required] The categorical prediction drift metric to calculate.
- threshold
Property Map
- The threshold value. If null, a default value will be set depending on the selected metric.
Classification, ClassificationArgs
Classification task in AutoML Table vertical.
ClassificationModels
Enum for all classification models supported by AutoML.
| Name | Value | Description |
|---|
| LogisticRegression | LogisticRegression | 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. |
| SGD | SGD | 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. |
| MultinomialNaiveBayes | MultinomialNaiveBayes | 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. |
| BernoulliNaiveBayes | BernoulliNaiveBayes | Naive Bayes classifier for multivariate Bernoulli models. |
| SVM | SVM | 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. |
| LinearSVM | LinearSVM | 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. |
| KNN | KNN | 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. |
| DecisionTree | DecisionTree | 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. |
| RandomForest | RandomForest | 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. |
| ExtremeRandomTrees | ExtremeRandomTrees | 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. |
| LightGBM | LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
| GradientBoosting | GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
| XGBoostClassifier | XGBoostClassifier | XGBoost: Extreme Gradient Boosting Algorithm. This algorithm is used for structured data where target column values can be divided into distinct class values. |
| Name | Value | Description |
|---|
| ClassificationModelsLogisticRegression | LogisticRegression | 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. |
| ClassificationModelsSGD | SGD | 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. |
| ClassificationModelsMultinomialNaiveBayes | MultinomialNaiveBayes | 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. |
| ClassificationModelsBernoulliNaiveBayes | BernoulliNaiveBayes | Naive Bayes classifier for multivariate Bernoulli models. |
| ClassificationModelsSVM | SVM | 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. |
| ClassificationModelsLinearSVM | LinearSVM | 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. |
| ClassificationModelsKNN | KNN | 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. |
| ClassificationModelsDecisionTree | DecisionTree | 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. |
| ClassificationModelsRandomForest | RandomForest | 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. |
| ClassificationModelsExtremeRandomTrees | ExtremeRandomTrees | 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. |
| ClassificationModelsLightGBM | LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
| ClassificationModelsGradientBoosting | GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
| ClassificationModelsXGBoostClassifier | XGBoostClassifier | XGBoost: Extreme Gradient Boosting Algorithm. This algorithm is used for structured data where target column values can be divided into distinct class values. |
| Value | Description |
|---|
LogisticRegression | 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. |
SGD | 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. |
MultinomialNaiveBayes | 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. |
BernoulliNaiveBayes | Naive Bayes classifier for multivariate Bernoulli models. |
SVM | 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. |
LinearSVM | 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. |
KNN | 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. |
DecisionTree | 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. |
RandomForest | 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. |
ExtremeRandomTrees | 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. |
LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
XGBoostClassifier | XGBoost: Extreme Gradient Boosting Algorithm. This algorithm is used for structured data where target column values can be divided into distinct class values. |
| Name | Value | Description |
|---|
| LogisticRegression | LogisticRegression | 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. |
| SGD | SGD | 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. |
| MultinomialNaiveBayes | MultinomialNaiveBayes | 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. |
| BernoulliNaiveBayes | BernoulliNaiveBayes | Naive Bayes classifier for multivariate Bernoulli models. |
| SVM | SVM | 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. |
| LinearSVM | LinearSVM | 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. |
| KNN | KNN | 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. |
| DecisionTree | DecisionTree | 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. |
| RandomForest | RandomForest | 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. |
| ExtremeRandomTrees | ExtremeRandomTrees | 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. |
| LightGBM | LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
| GradientBoosting | GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
| XGBoostClassifier | XGBoostClassifier | XGBoost: Extreme Gradient Boosting Algorithm. This algorithm is used for structured data where target column values can be divided into distinct class values. |
| Name | Value | Description |
|---|
| LogisticRegression | LogisticRegression | 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. |
| SGD | SGD | 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. |
| MultinomialNaiveBayes | MultinomialNaiveBayes | 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. |
| BernoulliNaiveBayes | BernoulliNaiveBayes | Naive Bayes classifier for multivariate Bernoulli models. |
| SVM | SVM | 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. |
| LinearSVM | LinearSVM | 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. |
| KNN | KNN | 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. |
| DecisionTree | DecisionTree | 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. |
| RandomForest | RandomForest | 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. |
| ExtremeRandomTrees | ExtremeRandomTrees | 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. |
| LightGBM | LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
| GradientBoosting | GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
| XGBoostClassifier | XGBoostClassifier | XGBoost: Extreme Gradient Boosting Algorithm. This algorithm is used for structured data where target column values can be divided into distinct class values. |
| Name | Value | Description |
|---|
| LOGISTIC_REGRESSION | LogisticRegression | 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. |
| SGD | SGD | 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. |
| MULTINOMIAL_NAIVE_BAYES | MultinomialNaiveBayes | 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. |
| BERNOULLI_NAIVE_BAYES | BernoulliNaiveBayes | Naive Bayes classifier for multivariate Bernoulli models. |
| SVM | SVM | 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 | LinearSVM | 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. |
| KNN | KNN | 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. |
| DECISION_TREE | DecisionTree | 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. |
| RANDOM_FOREST | RandomForest | 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. |
| EXTREME_RANDOM_TREES | ExtremeRandomTrees | 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. |
| LIGHT_GBM | LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
| GRADIENT_BOOSTING | GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
| XG_BOOST_CLASSIFIER | XGBoostClassifier | XGBoost: Extreme Gradient Boosting Algorithm. This algorithm is used for structured data where target column values can be divided into distinct class values. |
| Value | Description |
|---|
LogisticRegression | 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. |
SGD | 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. |
MultinomialNaiveBayes | 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. |
BernoulliNaiveBayes | Naive Bayes classifier for multivariate Bernoulli models. |
SVM | 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. |
LinearSVM | 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. |
KNN | 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. |
DecisionTree | 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. |
RandomForest | 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. |
ExtremeRandomTrees | 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. |
LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
XGBoostClassifier | 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 | Value | Description |
|---|
| AUCWeighted | AUCWeighted | 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 | Accuracy | Accuracy is the ratio of predictions that exactly match the true class labels. |
| NormMacroRecall | NormMacroRecall | 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. |
| AveragePrecisionScoreWeighted | AveragePrecisionScoreWeighted | The arithmetic mean of the average precision score for each class, weighted by
the number of true instances in each class. |
| PrecisionScoreWeighted | PrecisionScoreWeighted | The arithmetic mean of precision for each class, weighted by number of true instances in each class. |
| IOU | IOU | Intersection Over Union. Intersection of predictions divided by union of predictions. |
| Name | Value | Description |
|---|
| ClassificationMultilabelPrimaryMetricsAUCWeighted | AUCWeighted | 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. |
| ClassificationMultilabelPrimaryMetricsAccuracy | Accuracy | Accuracy is the ratio of predictions that exactly match the true class labels. |
| ClassificationMultilabelPrimaryMetricsNormMacroRecall | NormMacroRecall | 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. |
| ClassificationMultilabelPrimaryMetricsAveragePrecisionScoreWeighted | AveragePrecisionScoreWeighted | The arithmetic mean of the average precision score for each class, weighted by
the number of true instances in each class. |
| ClassificationMultilabelPrimaryMetricsPrecisionScoreWeighted | PrecisionScoreWeighted | The arithmetic mean of precision for each class, weighted by number of true instances in each class. |
| ClassificationMultilabelPrimaryMetricsIOU | IOU | Intersection Over Union. Intersection of predictions divided by union of predictions. |
| Value | Description |
|---|
AUCWeighted | 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 | Accuracy is the ratio of predictions that exactly match the true class labels. |
NormMacroRecall | 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. |
AveragePrecisionScoreWeighted | The arithmetic mean of the average precision score for each class, weighted by
the number of true instances in each class. |
PrecisionScoreWeighted | The arithmetic mean of precision for each class, weighted by number of true instances in each class. |
IOU | Intersection Over Union. Intersection of predictions divided by union of predictions. |
| Name | Value | Description |
|---|
| AUCWeighted | AUCWeighted | 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 | Accuracy | Accuracy is the ratio of predictions that exactly match the true class labels. |
| NormMacroRecall | NormMacroRecall | 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. |
| AveragePrecisionScoreWeighted | AveragePrecisionScoreWeighted | The arithmetic mean of the average precision score for each class, weighted by
the number of true instances in each class. |
| PrecisionScoreWeighted | PrecisionScoreWeighted | The arithmetic mean of precision for each class, weighted by number of true instances in each class. |
| IOU | IOU | Intersection Over Union. Intersection of predictions divided by union of predictions. |
| Name | Value | Description |
|---|
| AUCWeighted | AUCWeighted | 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 | Accuracy | Accuracy is the ratio of predictions that exactly match the true class labels. |
| NormMacroRecall | NormMacroRecall | 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. |
| AveragePrecisionScoreWeighted | AveragePrecisionScoreWeighted | The arithmetic mean of the average precision score for each class, weighted by
the number of true instances in each class. |
| PrecisionScoreWeighted | PrecisionScoreWeighted | The arithmetic mean of precision for each class, weighted by number of true instances in each class. |
| IOU | IOU | Intersection Over Union. Intersection of predictions divided by union of predictions. |
| Name | Value | Description |
|---|
| AUC_WEIGHTED | AUCWeighted | 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 | Accuracy | Accuracy is the ratio of predictions that exactly match the true class labels. |
| NORM_MACRO_RECALL | NormMacroRecall | 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. |
| AVERAGE_PRECISION_SCORE_WEIGHTED | AveragePrecisionScoreWeighted | The arithmetic mean of the average precision score for each class, weighted by
the number of true instances in each class. |
| PRECISION_SCORE_WEIGHTED | PrecisionScoreWeighted | The arithmetic mean of precision for each class, weighted by number of true instances in each class. |
| IOU | IOU | Intersection Over Union. Intersection of predictions divided by union of predictions. |
| Value | Description |
|---|
AUCWeighted | 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 | Accuracy is the ratio of predictions that exactly match the true class labels. |
NormMacroRecall | 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. |
AveragePrecisionScoreWeighted | The arithmetic mean of the average precision score for each class, weighted by
the number of true instances in each class. |
PrecisionScoreWeighted | The arithmetic mean of precision for each class, weighted by number of true instances in each class. |
IOU | Intersection Over Union. Intersection of predictions divided by union of predictions. |
ClassificationPrimaryMetrics
Primary metrics for classification tasks.
| Name | Value | Description |
|---|
| AUCWeighted | AUCWeighted | 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 | Accuracy | Accuracy is the ratio of predictions that exactly match the true class labels. |
| NormMacroRecall | NormMacroRecall | 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. |
| AveragePrecisionScoreWeighted | AveragePrecisionScoreWeighted | The arithmetic mean of the average precision score for each class, weighted by
the number of true instances in each class. |
| PrecisionScoreWeighted | PrecisionScoreWeighted | The arithmetic mean of precision for each class, weighted by number of true instances in each class. |
| Name | Value | Description |
|---|
| ClassificationPrimaryMetricsAUCWeighted | AUCWeighted | 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. |
| ClassificationPrimaryMetricsAccuracy | Accuracy | Accuracy is the ratio of predictions that exactly match the true class labels. |
| ClassificationPrimaryMetricsNormMacroRecall | NormMacroRecall | 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. |
| ClassificationPrimaryMetricsAveragePrecisionScoreWeighted | AveragePrecisionScoreWeighted | The arithmetic mean of the average precision score for each class, weighted by
the number of true instances in each class. |
| ClassificationPrimaryMetricsPrecisionScoreWeighted | PrecisionScoreWeighted | The arithmetic mean of precision for each class, weighted by number of true instances in each class. |
| Value | Description |
|---|
AUCWeighted | 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 | Accuracy is the ratio of predictions that exactly match the true class labels. |
NormMacroRecall | 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. |
AveragePrecisionScoreWeighted | The arithmetic mean of the average precision score for each class, weighted by
the number of true instances in each class. |
PrecisionScoreWeighted | The arithmetic mean of precision for each class, weighted by number of true instances in each class. |
| Name | Value | Description |
|---|
| AUCWeighted | AUCWeighted | 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 | Accuracy | Accuracy is the ratio of predictions that exactly match the true class labels. |
| NormMacroRecall | NormMacroRecall | 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. |
| AveragePrecisionScoreWeighted | AveragePrecisionScoreWeighted | The arithmetic mean of the average precision score for each class, weighted by
the number of true instances in each class. |
| PrecisionScoreWeighted | PrecisionScoreWeighted | The arithmetic mean of precision for each class, weighted by number of true instances in each class. |
| Name | Value | Description |
|---|
| AUCWeighted | AUCWeighted | 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 | Accuracy | Accuracy is the ratio of predictions that exactly match the true class labels. |
| NormMacroRecall | NormMacroRecall | 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. |
| AveragePrecisionScoreWeighted | AveragePrecisionScoreWeighted | The arithmetic mean of the average precision score for each class, weighted by
the number of true instances in each class. |
| PrecisionScoreWeighted | PrecisionScoreWeighted | The arithmetic mean of precision for each class, weighted by number of true instances in each class. |
| Name | Value | Description |
|---|
| AUC_WEIGHTED | AUCWeighted | 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 | Accuracy | Accuracy is the ratio of predictions that exactly match the true class labels. |
| NORM_MACRO_RECALL | NormMacroRecall | 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. |
| AVERAGE_PRECISION_SCORE_WEIGHTED | AveragePrecisionScoreWeighted | The arithmetic mean of the average precision score for each class, weighted by
the number of true instances in each class. |
| PRECISION_SCORE_WEIGHTED | PrecisionScoreWeighted | The arithmetic mean of precision for each class, weighted by number of true instances in each class. |
| Value | Description |
|---|
AUCWeighted | 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 | Accuracy is the ratio of predictions that exactly match the true class labels. |
NormMacroRecall | 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. |
AveragePrecisionScoreWeighted | The arithmetic mean of the average precision score for each class, weighted by
the number of true instances in each class. |
PrecisionScoreWeighted | The arithmetic mean of precision for each class, weighted by number of true instances in each class. |
ClassificationResponse
Classification task in AutoML Table vertical.
ClassificationTrainingSettings, ClassificationTrainingSettingsArgs
Classification Training related configuration.
- allowed_training_algorithms
list(string | "LogisticRegression" | "SGD" | "MultinomialNaiveBayes" | "BernoulliNaiveBayes" | "SVM" | "LinearSVM" | "KNN" | "DecisionTree" | "RandomForest" | "ExtremeRandomTrees" | "LightGBM" | "GradientBoosting" | "XGBoostClassifier")
- Allowed models for classification task.
- blocked_training_algorithms
list(string | "LogisticRegression" | "SGD" | "MultinomialNaiveBayes" | "BernoulliNaiveBayes" | "SVM" | "LinearSVM" | "KNN" | "DecisionTree" | "RandomForest" | "ExtremeRandomTrees" | "LightGBM" | "GradientBoosting" | "XGBoostClassifier")
- Blocked models for classification task.
- enable_dnn_training
bool
- Enable recommendation of DNN models.
- enable_model_explainability
bool
- Flag to turn on explainability on best model.
- enable_onnx_compatible_models
bool
- Flag for enabling onnx compatible models.
- enable_stack_ensemble
bool
- Enable stack ensemble run.
- enable_vote_ensemble
bool
- Enable voting ensemble run.
- ensemble_model_download_timeout
string
- 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.
- stack_ensemble_settings
object
- Stack ensemble settings for stack ensemble run.
- allowedTrainingAlgorithms
List<String | "LogisticRegression" | "SGD" | "MultinomialNaiveBayes" | "BernoulliNaiveBayes" | "SVM" | "LinearSVM" | "KNN" | "DecisionTree" | "RandomForest" | "ExtremeRandomTrees" | "LightGBM" | "GradientBoosting" | "XGBoostClassifier">
- Allowed models for classification task.
- blockedTrainingAlgorithms
List<String | "LogisticRegression" | "SGD" | "MultinomialNaiveBayes" | "BernoulliNaiveBayes" | "SVM" | "LinearSVM" | "KNN" | "DecisionTree" | "RandomForest" | "ExtremeRandomTrees" | "LightGBM" | "GradientBoosting" | "XGBoostClassifier">
- Blocked models for classification task.
- enableDnnTraining
Boolean
- Enable recommendation of DNN models.
- enableModelExplainability
Boolean
- Flag to turn on explainability on best model.
- enableOnnxCompatibleModels
Boolean
- Flag for enabling onnx compatible models.
- enableStackEnsemble
Boolean
- Enable stack ensemble run.
- enableVoteEnsemble
Boolean
- Enable voting ensemble run.
- ensembleModelDownloadTimeout
String
- 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
Property Map
- Stack ensemble settings for stack ensemble run.
ClassificationTrainingSettingsResponse
Classification Training related configuration.
Column transformer parameters.
- Fields
List<string>
- Fields to apply transformer logic on.
- Parameters
object
- Different properties to be passed to transformer.
Input expected is dictionary of key,value pairs in JSON format.
- Fields
[]string
- Fields to apply transformer logic on.
- Parameters
interface{}
- Different properties to be passed to transformer.
Input expected is dictionary of key,value pairs in JSON format.
- fields
list(string)
- Fields to apply transformer logic on.
- parameters
any
- Different properties to be passed to transformer.
Input expected is dictionary of key,value pairs in JSON format.
- fields
List<String>
- Fields to apply transformer logic on.
- parameters
Object
- Different properties to be passed to transformer.
Input expected is dictionary of key,value pairs in JSON format.
- fields
string[]
- Fields to apply transformer logic on.
- parameters
any
- Different properties to be passed to transformer.
Input expected is dictionary of key,value pairs in JSON format.
- fields
Sequence[str]
- Fields to apply transformer logic on.
- parameters
Any
- Different properties to be passed to transformer.
Input expected is dictionary of key,value pairs in JSON format.
- fields
List<String>
- Fields to apply transformer logic on.
- parameters
Any
- Different properties to be passed to transformer.
Input expected is dictionary of key,value pairs in JSON format.
Column transformer parameters.
- Fields
List<string>
- Fields to apply transformer logic on.
- Parameters
object
- Different properties to be passed to transformer.
Input expected is dictionary of key,value pairs in JSON format.
- Fields
[]string
- Fields to apply transformer logic on.
- Parameters
interface{}
- Different properties to be passed to transformer.
Input expected is dictionary of key,value pairs in JSON format.
- fields
list(string)
- Fields to apply transformer logic on.
- parameters
any
- Different properties to be passed to transformer.
Input expected is dictionary of key,value pairs in JSON format.
- fields
List<String>
- Fields to apply transformer logic on.
- parameters
Object
- Different properties to be passed to transformer.
Input expected is dictionary of key,value pairs in JSON format.
- fields
string[]
- Fields to apply transformer logic on.
- parameters
any
- Different properties to be passed to transformer.
Input expected is dictionary of key,value pairs in JSON format.
- fields
Sequence[str]
- Fields to apply transformer logic on.
- parameters
Any
- Different properties to be passed to transformer.
Input expected is dictionary of key,value pairs in JSON format.
- fields
List<String>
- Fields to apply transformer logic on.
- parameters
Any
- Different properties to be passed to transformer.
Input expected is dictionary of key,value pairs in JSON format.
CommandJob, CommandJobArgs
Command job definition.
- Command
string
- [Required] The command to execute on startup of the job. eg. "python train.py"
- EnvironmentId
string
- [Required] The ARM resource ID of the Environment specification for the job.
- CodeId
string
- ARM resource ID of the code asset.
- 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
Pulumi.AzureNative.MachineLearningServices.Inputs.Mpi | Pulumi.AzureNative.MachineLearningServices.Inputs.PyTorch | Pulumi.AzureNative.MachineLearningServices.Inputs.TensorFlow
- Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null.
- EnvironmentVariables
Dictionary<string, string>
- Environment variables included in the job.
- ExperimentName
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- Identity
Pulumi.AzureNative.MachineLearningServices.Inputs.AmlToken | Pulumi.AzureNative.MachineLearningServices.Inputs.ManagedIdentity | Pulumi.AzureNative.MachineLearningServices.Inputs.UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- Inputs
Dictionary<string, object>
- Mapping of input data bindings used in the job.
- IsArchived
bool
- Is the asset archived?
- Limits
Pulumi.AzureNative.MachineLearningServices.Inputs.CommandJobLimits
- Command Job limit.
- NotificationSetting
Pulumi.AzureNative.MachineLearningServices.Inputs.NotificationSetting
- Notification setting for the job
- Outputs
Dictionary<string, object>
- Mapping of output data bindings used in the job.
- Properties
Dictionary<string, string>
- The asset property dictionary.
- QueueSettings
Pulumi.AzureNative.MachineLearningServices.Inputs.QueueSettings
- Queue settings for the job
- Resources
Pulumi.AzureNative.MachineLearningServices.Inputs.JobResourceConfiguration
- Compute Resource configuration for the job.
- Services
Dictionary<string, Pulumi.AzureNative.MachineLearningServices.Inputs.JobService>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- Tags
Dictionary<string, string>
- Tag dictionary. Tags can be added, removed, and updated.
- Command
string
- [Required] The command to execute on startup of the job. eg. "python train.py"
- EnvironmentId
string
- [Required] The ARM resource ID of the Environment specification for the job.
- CodeId
string
- ARM resource ID of the code asset.
- 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
Mpi | PyTorch | TensorFlow
- Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null.
- EnvironmentVariables
map[string]string
- Environment variables included in the job.
- ExperimentName
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- Identity
AmlToken | ManagedIdentity | UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- Inputs
map[string]interface{}
- Mapping of input data bindings used in the job.
- IsArchived
bool
- Is the asset archived?
- Limits
CommandJobLimits
- Command Job limit.
- NotificationSetting
NotificationSetting
- Notification setting for the job
- Outputs
map[string]interface{}
- Mapping of output data bindings used in the job.
- Properties
map[string]string
- The asset property dictionary.
- QueueSettings
QueueSettings
- Queue settings for the job
- Resources
JobResourceConfiguration
- Compute Resource configuration for the job.
- Services
map[string]JobService
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- Tags
map[string]string
- Tag dictionary. Tags can be added, removed, and updated.
- command
string
- [Required] The command to execute on startup of the job. eg. "python train.py"
- environment_id
string
- [Required] The ARM resource ID of the Environment specification for the job.
- code_id
string
- ARM resource ID of the code asset.
- component_id
string
- ARM resource ID of the component resource.
- compute_id
string
- ARM resource ID of the compute resource.
- description
string
- The asset description text.
- display_name
string
- Display name of job.
- distribution
object | object | object
- Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null.
- environment_variables
map(string)
- Environment variables included in the job.
- experiment_name
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
object | object | object
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
map(object | object | object | object | object | object | object)
- Mapping of input data bindings used in the job.
- is_archived
bool
- Is the asset archived?
- limits
object
- Command Job limit.
- notification_setting
object
- Notification setting for the job
- outputs
map(object | object | object | object | object | object)
- Mapping of output data bindings used in the job.
- properties
map(string)
- The asset property dictionary.
- queue_settings
object
- Queue settings for the job
- resources
object
- Compute Resource configuration for the job.
- services
map(object)
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
map(string)
- Tag dictionary. Tags can be added, removed, and updated.
- command
String
- [Required] The command to execute on startup of the job. eg. "python train.py"
- environmentId
String
- [Required] The ARM resource ID of the Environment specification for the job.
- codeId
String
- ARM resource ID of the code asset.
- 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
Mpi | PyTorch | TensorFlow
- Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null.
- environmentVariables
Map<String,String>
- Environment variables included in the job.
- experimentName
String
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlToken | ManagedIdentity | UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Map<String,Object>
- Mapping of input data bindings used in the job.
- isArchived
Boolean
- Is the asset archived?
- limits
CommandJobLimits
- Command Job limit.
- notificationSetting
NotificationSetting
- Notification setting for the job
- outputs
Map<String,Object>
- Mapping of output data bindings used in the job.
- properties
Map<String,String>
- The asset property dictionary.
- queueSettings
QueueSettings
- Queue settings for the job
- resources
JobResourceConfiguration
- Compute Resource configuration for the job.
- services
Map<String,JobService>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Map<String,String>
- Tag dictionary. Tags can be added, removed, and updated.
- command
string
- [Required] The command to execute on startup of the job. eg. "python train.py"
- environmentId
string
- [Required] The ARM resource ID of the Environment specification for the job.
- codeId
string
- ARM resource ID of the code asset.
- 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
Mpi | PyTorch | TensorFlow
- Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null.
- environmentVariables
{[key: string]: string}
- Environment variables included in the job.
- experimentName
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlToken | ManagedIdentity | UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
{[key: string]: CustomModelJobInput | LiteralJobInput | MLFlowModelJobInput | MLTableJobInput | TritonModelJobInput | UriFileJobInput | UriFolderJobInput}
- Mapping of input data bindings used in the job.
- isArchived
boolean
- Is the asset archived?
- limits
CommandJobLimits
- Command Job limit.
- notificationSetting
NotificationSetting
- Notification setting for the job
- outputs
{[key: string]: CustomModelJobOutput | MLFlowModelJobOutput | MLTableJobOutput | TritonModelJobOutput | UriFileJobOutput | UriFolderJobOutput}
- Mapping of output data bindings used in the job.
- properties
{[key: string]: string}
- The asset property dictionary.
- queueSettings
QueueSettings
- Queue settings for the job
- resources
JobResourceConfiguration
- Compute Resource configuration for the job.
- services
{[key: string]: JobService}
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
{[key: string]: string}
- Tag dictionary. Tags can be added, removed, and updated.
- command
str
- [Required] The command to execute on startup of the job. eg. "python train.py"
- environment_id
str
- [Required] The ARM resource ID of the Environment specification for the job.
- code_id
str
- ARM resource ID of the code asset.
- component_id
str
- ARM resource ID of the component resource.
- compute_id
str
- ARM resource ID of the compute resource.
- description
str
- The asset description text.
- display_name
str
- Display name of job.
- distribution
Mpi | PyTorch | TensorFlow
- Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null.
- environment_variables
Mapping[str, str]
- Environment variables included in the job.
- experiment_name
str
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlToken | ManagedIdentity | UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Mapping[str, Union[CustomModelJobInput, LiteralJobInput, MLFlowModelJobInput, MLTableJobInput, TritonModelJobInput, UriFileJobInput, UriFolderJobInput]]
- Mapping of input data bindings used in the job.
- is_archived
bool
- Is the asset archived?
- limits
CommandJobLimits
- Command Job limit.
- notification_setting
NotificationSetting
- Notification setting for the job
- outputs
Mapping[str, Union[CustomModelJobOutput, MLFlowModelJobOutput, MLTableJobOutput, TritonModelJobOutput, UriFileJobOutput, UriFolderJobOutput]]
- Mapping of output data bindings used in the job.
- properties
Mapping[str, str]
- The asset property dictionary.
- queue_settings
QueueSettings
- Queue settings for the job
- resources
JobResourceConfiguration
- Compute Resource configuration for the job.
- services
Mapping[str, JobService]
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Mapping[str, str]
- Tag dictionary. Tags can be added, removed, and updated.
- command
String
- [Required] The command to execute on startup of the job. eg. "python train.py"
- environmentId
String
- [Required] The ARM resource ID of the Environment specification for the job.
- codeId
String
- ARM resource ID of the code asset.
- 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
Property Map | Property Map | Property Map
- Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null.
- environmentVariables
Map<String>
- Environment variables included in the job.
- experimentName
String
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
Property Map | Property Map | Property Map
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Map<Property Map | Property Map | Property Map | Property Map | Property Map | Property Map | Property Map>
- Mapping of input data bindings used in the job.
- isArchived
Boolean
- Is the asset archived?
- limits
Property Map
- Command Job limit.
- notificationSetting
Property Map
- Notification setting for the job
- outputs
Map<Property Map | Property Map | Property Map | Property Map | Property Map | Property Map>
- Mapping of output data bindings used in the job.
- properties
Map<String>
- The asset property dictionary.
- queueSettings
Property Map
- Queue settings for the job
- resources
Property Map
- Compute Resource configuration for the job.
- services
Map<Property Map>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Map<String>
- Tag dictionary. Tags can be added, removed, and updated.
CommandJobLimits, CommandJobLimitsArgs
Command Job limit class.
- 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.
- 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.
- 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.
- 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.
- 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.
- timeout
str
- The max run duration in ISO 8601 format, after which the job will be cancelled. Only supports duration with precision as low as Seconds.
- 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.
CommandJobLimitsResponse
Command Job limit class.
- 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.
- 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.
- 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.
- 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.
- 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.
- timeout
str
- The max run duration in ISO 8601 format, after which the job will be cancelled. Only supports duration with precision as low as Seconds.
- 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.
CommandJobResponse
Command job definition.
- Command
string
- [Required] The command to execute on startup of the job. eg. "python train.py"
- EnvironmentId
string
- [Required] The ARM resource ID of the Environment specification for the job.
- Parameters
object
- Input parameters.
- Status
string
- Status of the job.
- CodeId
string
- ARM resource ID of the code asset.
- 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
Pulumi.AzureNative.MachineLearningServices.Inputs.MpiResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.PyTorchResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.TensorFlowResponse
- Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null.
- EnvironmentVariables
Dictionary<string, string>
- Environment variables included in the job.
- ExperimentName
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- Identity
Pulumi.AzureNative.MachineLearningServices.Inputs.AmlTokenResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.ManagedIdentityResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- Inputs
Dictionary<string, object>
- Mapping of input data bindings used in the job.
- IsArchived
bool
- Is the asset archived?
- Limits
Pulumi.AzureNative.MachineLearningServices.Inputs.CommandJobLimitsResponse
- Command Job limit.
- NotificationSetting
Pulumi.AzureNative.MachineLearningServices.Inputs.NotificationSettingResponse
- Notification setting for the job
- Outputs
Dictionary<string, object>
- Mapping of output data bindings used in the job.
- Properties
Dictionary<string, string>
- The asset property dictionary.
- QueueSettings
Pulumi.AzureNative.MachineLearningServices.Inputs.QueueSettingsResponse
- Queue settings for the job
- Resources
Pulumi.AzureNative.MachineLearningServices.Inputs.JobResourceConfigurationResponse
- Compute Resource configuration for the job.
- Services
Dictionary<string, Pulumi.AzureNative.MachineLearningServices.Inputs.JobServiceResponse>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- Tags
Dictionary<string, string>
- Tag dictionary. Tags can be added, removed, and updated.
- Command
string
- [Required] The command to execute on startup of the job. eg. "python train.py"
- EnvironmentId
string
- [Required] The ARM resource ID of the Environment specification for the job.
- Parameters
interface{}
- Input parameters.
- Status
string
- Status of the job.
- CodeId
string
- ARM resource ID of the code asset.
- 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
MpiResponse | PyTorchResponse | TensorFlowResponse
- Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null.
- EnvironmentVariables
map[string]string
- Environment variables included in the job.
- ExperimentName
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- Identity
AmlTokenResponse | ManagedIdentityResponse | UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- Inputs
map[string]interface{}
- Mapping of input data bindings used in the job.
- IsArchived
bool
- Is the asset archived?
- Limits
CommandJobLimitsResponse
- Command Job limit.
- NotificationSetting
NotificationSettingResponse
- Notification setting for the job
- Outputs
map[string]interface{}
- Mapping of output data bindings used in the job.
- Properties
map[string]string
- The asset property dictionary.
- QueueSettings
QueueSettingsResponse
- Queue settings for the job
- Resources
JobResourceConfigurationResponse
- Compute Resource configuration for the job.
- Services
map[string]JobServiceResponse
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- Tags
map[string]string
- Tag dictionary. Tags can be added, removed, and updated.
- command
string
- [Required] The command to execute on startup of the job. eg. "python train.py"
- environment_id
string
- [Required] The ARM resource ID of the Environment specification for the job.
- parameters
any
- Input parameters.
- status
string
- Status of the job.
- code_id
string
- ARM resource ID of the code asset.
- component_id
string
- ARM resource ID of the component resource.
- compute_id
string
- ARM resource ID of the compute resource.
- description
string
- The asset description text.
- display_name
string
- Display name of job.
- distribution
object | object | object
- Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null.
- environment_variables
map(string)
- Environment variables included in the job.
- experiment_name
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
object | object | object
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
map(object | object | object | object | object | object | object)
- Mapping of input data bindings used in the job.
- is_archived
bool
- Is the asset archived?
- limits
object
- Command Job limit.
- notification_setting
object
- Notification setting for the job
- outputs
map(object | object | object | object | object | object)
- Mapping of output data bindings used in the job.
- properties
map(string)
- The asset property dictionary.
- queue_settings
object
- Queue settings for the job
- resources
object
- Compute Resource configuration for the job.
- services
map(object)
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
map(string)
- Tag dictionary. Tags can be added, removed, and updated.
- command
String
- [Required] The command to execute on startup of the job. eg. "python train.py"
- environmentId
String
- [Required] The ARM resource ID of the Environment specification for the job.
- parameters
Object
- Input parameters.
- status
String
- Status of the job.
- codeId
String
- ARM resource ID of the code asset.
- 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
MpiResponse | PyTorchResponse | TensorFlowResponse
- Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null.
- environmentVariables
Map<String,String>
- Environment variables included in the job.
- experimentName
String
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlTokenResponse | ManagedIdentityResponse | UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Map<String,Object>
- Mapping of input data bindings used in the job.
- isArchived
Boolean
- Is the asset archived?
- limits
CommandJobLimitsResponse
- Command Job limit.
- notificationSetting
NotificationSettingResponse
- Notification setting for the job
- outputs
Map<String,Object>
- Mapping of output data bindings used in the job.
- properties
Map<String,String>
- The asset property dictionary.
- queueSettings
QueueSettingsResponse
- Queue settings for the job
- resources
JobResourceConfigurationResponse
- Compute Resource configuration for the job.
- services
Map<String,JobServiceResponse>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Map<String,String>
- Tag dictionary. Tags can be added, removed, and updated.
- command
string
- [Required] The command to execute on startup of the job. eg. "python train.py"
- environmentId
string
- [Required] The ARM resource ID of the Environment specification for the job.
- parameters
any
- Input parameters.
- status
string
- Status of the job.
- codeId
string
- ARM resource ID of the code asset.
- 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
MpiResponse | PyTorchResponse | TensorFlowResponse
- Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null.
- environmentVariables
{[key: string]: string}
- Environment variables included in the job.
- experimentName
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlTokenResponse | ManagedIdentityResponse | UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
{[key: string]: CustomModelJobInputResponse | LiteralJobInputResponse | MLFlowModelJobInputResponse | MLTableJobInputResponse | TritonModelJobInputResponse | UriFileJobInputResponse | UriFolderJobInputResponse}
- Mapping of input data bindings used in the job.
- isArchived
boolean
- Is the asset archived?
- limits
CommandJobLimitsResponse
- Command Job limit.
- notificationSetting
NotificationSettingResponse
- Notification setting for the job
- outputs
{[key: string]: CustomModelJobOutputResponse | MLFlowModelJobOutputResponse | MLTableJobOutputResponse | TritonModelJobOutputResponse | UriFileJobOutputResponse | UriFolderJobOutputResponse}
- Mapping of output data bindings used in the job.
- properties
{[key: string]: string}
- The asset property dictionary.
- queueSettings
QueueSettingsResponse
- Queue settings for the job
- resources
JobResourceConfigurationResponse
- Compute Resource configuration for the job.
- services
{[key: string]: JobServiceResponse}
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
{[key: string]: string}
- Tag dictionary. Tags can be added, removed, and updated.
- command
str
- [Required] The command to execute on startup of the job. eg. "python train.py"
- environment_id
str
- [Required] The ARM resource ID of the Environment specification for the job.
- parameters
Any
- Input parameters.
- status
str
- Status of the job.
- code_id
str
- ARM resource ID of the code asset.
- component_id
str
- ARM resource ID of the component resource.
- compute_id
str
- ARM resource ID of the compute resource.
- description
str
- The asset description text.
- display_name
str
- Display name of job.
- distribution
MpiResponse | PyTorchResponse | TensorFlowResponse
- Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null.
- environment_variables
Mapping[str, str]
- Environment variables included in the job.
- experiment_name
str
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlTokenResponse | ManagedIdentityResponse | UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Mapping[str, Union[CustomModelJobInputResponse, LiteralJobInputResponse, MLFlowModelJobInputResponse, MLTableJobInputResponse, TritonModelJobInputResponse, UriFileJobInputResponse, UriFolderJobInputResponse]]
- Mapping of input data bindings used in the job.
- is_archived
bool
- Is the asset archived?
- limits
CommandJobLimitsResponse
- Command Job limit.
- notification_setting
NotificationSettingResponse
- Notification setting for the job
- outputs
Mapping[str, Union[CustomModelJobOutputResponse, MLFlowModelJobOutputResponse, MLTableJobOutputResponse, TritonModelJobOutputResponse, UriFileJobOutputResponse, UriFolderJobOutputResponse]]
- Mapping of output data bindings used in the job.
- properties
Mapping[str, str]
- The asset property dictionary.
- queue_settings
QueueSettingsResponse
- Queue settings for the job
- resources
JobResourceConfigurationResponse
- Compute Resource configuration for the job.
- services
Mapping[str, JobServiceResponse]
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Mapping[str, str]
- Tag dictionary. Tags can be added, removed, and updated.
- command
String
- [Required] The command to execute on startup of the job. eg. "python train.py"
- environmentId
String
- [Required] The ARM resource ID of the Environment specification for the job.
- parameters
Any
- Input parameters.
- status
String
- Status of the job.
- codeId
String
- ARM resource ID of the code asset.
- 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
Property Map | Property Map | Property Map
- Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null.
- environmentVariables
Map<String>
- Environment variables included in the job.
- experimentName
String
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
Property Map | Property Map | Property Map
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Map<Property Map | Property Map | Property Map | Property Map | Property Map | Property Map | Property Map>
- Mapping of input data bindings used in the job.
- isArchived
Boolean
- Is the asset archived?
- limits
Property Map
- Command Job limit.
- notificationSetting
Property Map
- Notification setting for the job
- outputs
Map<Property Map | Property Map | Property Map | Property Map | Property Map | Property Map>
- Mapping of output data bindings used in the job.
- properties
Map<String>
- The asset property dictionary.
- queueSettings
Property Map
- Queue settings for the job
- resources
Property Map
- Compute Resource configuration for the job.
- services
Map<Property Map>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Map<String>
- Tag dictionary. Tags can be added, removed, and updated.
CreateMonitorAction, CreateMonitorActionArgs
CreateMonitorActionResponse
CronTrigger, CronTriggerArgs
CronTriggerResponse
CustomForecastHorizon, CustomForecastHorizonArgs
The desired maximum forecast horizon in units of time-series frequency.
- Value
int
- [Required] Forecast horizon value.
- Value
int
- [Required] Forecast horizon value.
- value
number
- [Required] Forecast horizon value.
- value
Integer
- [Required] Forecast horizon value.
- value
number
- [Required] Forecast horizon value.
- value
int
- [Required] Forecast horizon value.
- value
Number
- [Required] Forecast horizon value.
CustomForecastHorizonResponse
The desired maximum forecast horizon in units of time-series frequency.
- Value
int
- [Required] Forecast horizon value.
- Value
int
- [Required] Forecast horizon value.
- value
number
- [Required] Forecast horizon value.
- value
Integer
- [Required] Forecast horizon value.
- value
number
- [Required] Forecast horizon value.
- value
int
- [Required] Forecast horizon value.
- value
Number
- [Required] Forecast horizon value.
CustomMetricThreshold, CustomMetricThresholdArgs
- Metric
string
- [Required] The user-defined metric to calculate.
- Threshold
MonitoringThreshold
- The threshold value. If null, a default value will be set depending on the selected metric.
- metric
string
- [Required] The user-defined metric to calculate.
- threshold
object
- The threshold value. If null, a default value will be set depending on the selected metric.
- metric
String
- [Required] The user-defined metric to calculate.
- threshold
MonitoringThreshold
- The threshold value. If null, a default value will be set depending on the selected metric.
- metric
string
- [Required] The user-defined metric to calculate.
- threshold
MonitoringThreshold
- The threshold value. If null, a default value will be set depending on the selected metric.
- metric
str
- [Required] The user-defined metric to calculate.
- threshold
MonitoringThreshold
- The threshold value. If null, a default value will be set depending on the selected metric.
- metric
String
- [Required] The user-defined metric to calculate.
- threshold
Property Map
- The threshold value. If null, a default value will be set depending on the selected metric.
CustomMetricThresholdResponse
- metric
string
- [Required] The user-defined metric to calculate.
- threshold
object
- The threshold value. If null, a default value will be set depending on the selected metric.
- metric
String
- [Required] The user-defined metric to calculate.
- threshold
Property Map
- The threshold value. If null, a default value will be set depending on the selected metric.
- Uri
string
- [Required] Input Asset URI.
- Description
string
- Description for the input.
- Mode
string
- Enum to determine the input data delivery mode.
- Uri
string
- [Required] Input Asset URI.
- Description
string
- Description for the input.
- Mode
string
- Enum to determine the input data delivery mode.
- uri
string
- [Required] Input Asset URI.
- description
string
- Description for the input.
- mode
string
- Enum to determine the input data delivery mode.
- uri
String
- [Required] Input Asset URI.
- description
String
- Description for the input.
- mode
String
- Enum to determine the input data delivery mode.
- uri
string
- [Required] Input Asset URI.
- description
string
- Description for the input.
- mode
string
- Enum to determine the input data delivery mode.
- uri
str
- [Required] Input Asset URI.
- description
str
- Description for the input.
- mode
str
- Enum to determine the input data delivery mode.
- uri
String
- [Required] Input Asset URI.
- description
String
- Description for the input.
- mode
String
- Enum to determine the input data delivery mode.
CustomModelJobOutput, CustomModelJobOutputArgs
CustomModelJobOutputResponse
- AssetName
string
- Output Asset Name.
- Description
string
- Description for the output.
- Mode
string
- Output data delivery mode enums.
- Uri
string
- Output Asset URI.
- AssetName
string
- Output Asset Name.
- Description
string
- Description for the output.
- Mode
string
- Output data delivery mode enums.
- Uri
string
- Output Asset URI.
- asset_name
string
- Output Asset Name.
- description
string
- Description for the output.
- mode
string
- Output data delivery mode enums.
- uri
string
- Output Asset URI.
- assetName
String
- Output Asset Name.
- description
String
- Description for the output.
- mode
String
- Output data delivery mode enums.
- uri
String
- Output Asset URI.
- assetName
string
- Output Asset Name.
- description
string
- Description for the output.
- mode
string
- Output data delivery mode enums.
- uri
string
- Output Asset URI.
- assetName
String
- Output Asset Name.
- description
String
- Description for the output.
- mode
String
- Output data delivery mode enums.
- uri
String
- Output Asset URI.
CustomMonitoringSignal, CustomMonitoringSignalArgs
- ComponentId
string
- [Required] Reference to the component asset used to calculate the custom metrics.
- MetricThresholds
List<Pulumi.AzureNative.MachineLearningServices.Inputs.CustomMetricThreshold>
- [Required] A list of metrics to calculate and their associated thresholds.
- InputAssets
Dictionary<string, object>
- Monitoring assets to take as input. Key is the component input port name, value is the data asset.
- Inputs
Dictionary<string, object>
- Extra component parameters to take as input. Key is the component literal input port name, value is the parameter value.
- NotificationTypes
List<Union<string, Pulumi.AzureNative.MachineLearningServices.MonitoringNotificationType>>
- The current notification mode for this signal.
- Properties
Dictionary<string, string>
- Property dictionary. Properties can be added, but not removed or altered.
- ComponentId
string
- [Required] Reference to the component asset used to calculate the custom metrics.
- MetricThresholds
[]CustomMetricThreshold
- [Required] A list of metrics to calculate and their associated thresholds.
- InputAssets
map[string]interface{}
- Monitoring assets to take as input. Key is the component input port name, value is the data asset.
- Inputs
map[string]interface{}
- Extra component parameters to take as input. Key is the component literal input port name, value is the parameter value.
- NotificationTypes
[]string
- The current notification mode for this signal.
- Properties
map[string]string
- Property dictionary. Properties can be added, but not removed or altered.
- component_id
string
- [Required] Reference to the component asset used to calculate the custom metrics.
- metric_thresholds
list(object)
- [Required] A list of metrics to calculate and their associated thresholds.
- input_assets
map(object | object | object)
- Monitoring assets to take as input. Key is the component input port name, value is the data asset.
- inputs
map(object | object | object | object | object | object | object)
- Extra component parameters to take as input. Key is the component literal input port name, value is the parameter value.
- notification_types
list(string | "AmlNotification")
- The current notification mode for this signal.
- properties
map(string)
- Property dictionary. Properties can be added, but not removed or altered.
- componentId
String
- [Required] Reference to the component asset used to calculate the custom metrics.
- metricThresholds
List<CustomMetricThreshold>
- [Required] A list of metrics to calculate and their associated thresholds.
- inputAssets
Map<String,Object>
- Monitoring assets to take as input. Key is the component input port name, value is the data asset.
- inputs
Map<String,Object>
- Extra component parameters to take as input. Key is the component literal input port name, value is the parameter value.
- notificationTypes
List<Either<String,MonitoringNotificationType>>
- The current notification mode for this signal.
- properties
Map<String,String>
- Property dictionary. Properties can be added, but not removed or altered.
- componentId
string
- [Required] Reference to the component asset used to calculate the custom metrics.
- metricThresholds
CustomMetricThreshold[]
- [Required] A list of metrics to calculate and their associated thresholds.
- inputAssets
{[key: string]: FixedInputData | RollingInputData | StaticInputData}
- Monitoring assets to take as input. Key is the component input port name, value is the data asset.
- inputs
{[key: string]: CustomModelJobInput | LiteralJobInput | MLFlowModelJobInput | MLTableJobInput | TritonModelJobInput | UriFileJobInput | UriFolderJobInput}
- Extra component parameters to take as input. Key is the component literal input port name, value is the parameter value.
- notificationTypes
(string | MonitoringNotificationType)[]
- The current notification mode for this signal.
- properties
{[key: string]: string}
- Property dictionary. Properties can be added, but not removed or altered.
- component_id
str
- [Required] Reference to the component asset used to calculate the custom metrics.
- metric_thresholds
Sequence[CustomMetricThreshold]
- [Required] A list of metrics to calculate and their associated thresholds.
- input_assets
Mapping[str, Union[FixedInputData, RollingInputData, StaticInputData]]
- Monitoring assets to take as input. Key is the component input port name, value is the data asset.
- inputs
Mapping[str, Union[CustomModelJobInput, LiteralJobInput, MLFlowModelJobInput, MLTableJobInput, TritonModelJobInput, UriFileJobInput, UriFolderJobInput]]
- Extra component parameters to take as input. Key is the component literal input port name, value is the parameter value.
- notification_types
Sequence[Union[str, MonitoringNotificationType]]
- The current notification mode for this signal.
- properties
Mapping[str, str]
- Property dictionary. Properties can be added, but not removed or altered.
- componentId
String
- [Required] Reference to the component asset used to calculate the custom metrics.
- metricThresholds
List<Property Map>
- [Required] A list of metrics to calculate and their associated thresholds.
- inputAssets
Map<Property Map | Property Map | Property Map>
- Monitoring assets to take as input. Key is the component input port name, value is the data asset.
- inputs
Map<Property Map | Property Map | Property Map | Property Map | Property Map | Property Map | Property Map>
- Extra component parameters to take as input. Key is the component literal input port name, value is the parameter value.
- notificationTypes
List<String | "AmlNotification">
- The current notification mode for this signal.
- properties
Map<String>
- Property dictionary. Properties can be added, but not removed or altered.
CustomMonitoringSignalResponse
- ComponentId
string
- [Required] Reference to the component asset used to calculate the custom metrics.
- MetricThresholds
[]CustomMetricThresholdResponse
- [Required] A list of metrics to calculate and their associated thresholds.
- InputAssets
map[string]interface{}
- Monitoring assets to take as input. Key is the component input port name, value is the data asset.
- Inputs
map[string]interface{}
- Extra component parameters to take as input. Key is the component literal input port name, value is the parameter value.
- NotificationTypes
[]string
- The current notification mode for this signal.
- Properties
map[string]string
- Property dictionary. Properties can be added, but not removed or altered.
- component_id
string
- [Required] Reference to the component asset used to calculate the custom metrics.
- metric_thresholds
list(object)
- [Required] A list of metrics to calculate and their associated thresholds.
- input_assets
map(object | object | object)
- Monitoring assets to take as input. Key is the component input port name, value is the data asset.
- inputs
map(object | object | object | object | object | object | object)
- Extra component parameters to take as input. Key is the component literal input port name, value is the parameter value.
- notification_types
list(string)
- The current notification mode for this signal.
- properties
map(string)
- Property dictionary. Properties can be added, but not removed or altered.
- componentId
String
- [Required] Reference to the component asset used to calculate the custom metrics.
- metricThresholds
List<CustomMetricThresholdResponse>
- [Required] A list of metrics to calculate and their associated thresholds.
- inputAssets
Map<String,Object>
- Monitoring assets to take as input. Key is the component input port name, value is the data asset.
- inputs
Map<String,Object>
- Extra component parameters to take as input. Key is the component literal input port name, value is the parameter value.
- notificationTypes
List<String>
- The current notification mode for this signal.
- properties
Map<String,String>
- Property dictionary. Properties can be added, but not removed or altered.
- componentId
string
- [Required] Reference to the component asset used to calculate the custom metrics.
- metricThresholds
CustomMetricThresholdResponse[]
- [Required] A list of metrics to calculate and their associated thresholds.
- inputAssets
{[key: string]: FixedInputDataResponse | RollingInputDataResponse | StaticInputDataResponse}
- Monitoring assets to take as input. Key is the component input port name, value is the data asset.
- inputs
{[key: string]: CustomModelJobInputResponse | LiteralJobInputResponse | MLFlowModelJobInputResponse | MLTableJobInputResponse | TritonModelJobInputResponse | UriFileJobInputResponse | UriFolderJobInputResponse}
- Extra component parameters to take as input. Key is the component literal input port name, value is the parameter value.
- notificationTypes
string[]
- The current notification mode for this signal.
- properties
{[key: string]: string}
- Property dictionary. Properties can be added, but not removed or altered.
- component_id
str
- [Required] Reference to the component asset used to calculate the custom metrics.
- metric_thresholds
Sequence[CustomMetricThresholdResponse]
- [Required] A list of metrics to calculate and their associated thresholds.
- input_assets
Mapping[str, Union[FixedInputDataResponse, RollingInputDataResponse, StaticInputDataResponse]]
- Monitoring assets to take as input. Key is the component input port name, value is the data asset.
- inputs
Mapping[str, Union[CustomModelJobInputResponse, LiteralJobInputResponse, MLFlowModelJobInputResponse, MLTableJobInputResponse, TritonModelJobInputResponse, UriFileJobInputResponse, UriFolderJobInputResponse]]
- Extra component parameters to take as input. Key is the component literal input port name, value is the parameter value.
- notification_types
Sequence[str]
- The current notification mode for this signal.
- properties
Mapping[str, str]
- Property dictionary. Properties can be added, but not removed or altered.
- componentId
String
- [Required] Reference to the component asset used to calculate the custom metrics.
- metricThresholds
List<Property Map>
- [Required] A list of metrics to calculate and their associated thresholds.
- inputAssets
Map<Property Map | Property Map | Property Map>
- Monitoring assets to take as input. Key is the component input port name, value is the data asset.
- inputs
Map<Property Map | Property Map | Property Map | Property Map | Property Map | Property Map | Property Map>
- Extra component parameters to take as input. Key is the component literal input port name, value is the parameter value.
- notificationTypes
List<String>
- The current notification mode for this signal.
- properties
Map<String>
- Property dictionary. Properties can be added, but not removed or altered.
CustomNCrossValidations, CustomNCrossValidationsArgs
N-Cross validations are specified by user.
- Value
int
- [Required] N-Cross validations value.
- Value
int
- [Required] N-Cross validations value.
- value
number
- [Required] N-Cross validations value.
- value
Integer
- [Required] N-Cross validations value.
- value
number
- [Required] N-Cross validations value.
- value
int
- [Required] N-Cross validations value.
- value
Number
- [Required] N-Cross validations value.
CustomNCrossValidationsResponse
N-Cross validations are specified by user.
- Value
int
- [Required] N-Cross validations value.
- Value
int
- [Required] N-Cross validations value.
- value
number
- [Required] N-Cross validations value.
- value
Integer
- [Required] N-Cross validations value.
- value
number
- [Required] N-Cross validations value.
- value
int
- [Required] N-Cross validations value.
- value
Number
- [Required] N-Cross validations value.
CustomSeasonality, CustomSeasonalityArgs
- Value
int
- [Required] Seasonality value.
- Value
int
- [Required] Seasonality value.
- value
number
- [Required] Seasonality value.
- value
Integer
- [Required] Seasonality value.
- value
number
- [Required] Seasonality value.
- value
int
- [Required] Seasonality value.
- value
Number
- [Required] Seasonality value.
CustomSeasonalityResponse
- Value
int
- [Required] Seasonality value.
- Value
int
- [Required] Seasonality value.
- value
number
- [Required] Seasonality value.
- value
Integer
- [Required] Seasonality value.
- value
number
- [Required] Seasonality value.
- value
int
- [Required] Seasonality value.
- value
Number
- [Required] Seasonality value.
CustomTargetLags, CustomTargetLagsArgs
- Values
List<int>
- [Required] Set target lags values.
- Values
[]int
- [Required] Set target lags values.
- values
list(number)
- [Required] Set target lags values.
- values
List<Integer>
- [Required] Set target lags values.
- values
number[]
- [Required] Set target lags values.
- values
Sequence[int]
- [Required] Set target lags values.
- values
List<Number>
- [Required] Set target lags values.
CustomTargetLagsResponse
- Values
List<int>
- [Required] Set target lags values.
- Values
[]int
- [Required] Set target lags values.
- values
list(number)
- [Required] Set target lags values.
- values
List<Integer>
- [Required] Set target lags values.
- values
number[]
- [Required] Set target lags values.
- values
Sequence[int]
- [Required] Set target lags values.
- values
List<Number>
- [Required] Set target lags values.
CustomTargetRollingWindowSize, CustomTargetRollingWindowSizeArgs
- Value
int
- [Required] TargetRollingWindowSize value.
- Value
int
- [Required] TargetRollingWindowSize value.
- value
number
- [Required] TargetRollingWindowSize value.
- value
Integer
- [Required] TargetRollingWindowSize value.
- value
number
- [Required] TargetRollingWindowSize value.
- value
int
- [Required] TargetRollingWindowSize value.
- value
Number
- [Required] TargetRollingWindowSize value.
CustomTargetRollingWindowSizeResponse
- Value
int
- [Required] TargetRollingWindowSize value.
- Value
int
- [Required] TargetRollingWindowSize value.
- value
number
- [Required] TargetRollingWindowSize value.
- value
Integer
- [Required] TargetRollingWindowSize value.
- value
number
- [Required] TargetRollingWindowSize value.
- value
int
- [Required] TargetRollingWindowSize value.
- value
Number
- [Required] TargetRollingWindowSize value.
DataDriftMonitoringSignal, DataDriftMonitoringSignalArgs
DataDriftMonitoringSignalResponse
DataQualityMonitoringSignal, DataQualityMonitoringSignalArgs
DataQualityMonitoringSignalResponse
EmailNotificationEnableType
Enum to determine the email notification type.
| Name | Value |
|---|
| JobCompleted | JobCompleted |
| JobFailed | JobFailed |
| JobCancelled | JobCancelled |
| Name | Value |
|---|
| EmailNotificationEnableTypeJobCompleted | JobCompleted |
| EmailNotificationEnableTypeJobFailed | JobFailed |
| EmailNotificationEnableTypeJobCancelled | JobCancelled |
| Value |
|---|
JobCompleted |
JobFailed |
JobCancelled |
| Name | Value |
|---|
| JobCompleted | JobCompleted |
| JobFailed | JobFailed |
| JobCancelled | JobCancelled |
| Name | Value |
|---|
| JobCompleted | JobCompleted |
| JobFailed | JobFailed |
| JobCancelled | JobCancelled |
| Name | Value |
|---|
| JOB_COMPLETED | JobCompleted |
| JOB_FAILED | JobFailed |
| JOB_CANCELLED | JobCancelled |
| Value |
|---|
JobCompleted |
JobFailed |
JobCancelled |
EndpointScheduleAction, EndpointScheduleActionArgs
EndpointScheduleActionResponse
FeatureAttributionDriftMonitoringSignal, FeatureAttributionDriftMonitoringSignalArgs
FeatureAttributionDriftMonitoringSignalResponse
FeatureAttributionMetric
[Required] The feature attribution metric to calculate.
| Name | Value | Description |
|---|
| NormalizedDiscountedCumulativeGain | NormalizedDiscountedCumulativeGain | The Normalized Discounted Cumulative Gain metric. |
| Name | Value | Description |
|---|
| FeatureAttributionMetricNormalizedDiscountedCumulativeGain | NormalizedDiscountedCumulativeGain | The Normalized Discounted Cumulative Gain metric. |
| Value | Description |
|---|
NormalizedDiscountedCumulativeGain | The Normalized Discounted Cumulative Gain metric. |
| Name | Value | Description |
|---|
| NormalizedDiscountedCumulativeGain | NormalizedDiscountedCumulativeGain | The Normalized Discounted Cumulative Gain metric. |
| Name | Value | Description |
|---|
| NormalizedDiscountedCumulativeGain | NormalizedDiscountedCumulativeGain | The Normalized Discounted Cumulative Gain metric. |
| Name | Value | Description |
|---|
| NORMALIZED_DISCOUNTED_CUMULATIVE_GAIN | NormalizedDiscountedCumulativeGain | The Normalized Discounted Cumulative Gain metric. |
| Value | Description |
|---|
NormalizedDiscountedCumulativeGain | The Normalized Discounted Cumulative Gain metric. |
FeatureAttributionMetricThreshold, FeatureAttributionMetricThresholdArgs
FeatureAttributionMetricThresholdResponse
- metric
string
- [Required] The feature attribution metric to calculate.
- threshold
object
- The threshold value. If null, a default value will be set depending on the selected metric.
- metric
String
- [Required] The feature attribution metric to calculate.
- threshold
Property Map
- 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 | Value | Description |
|---|
| Disabled | Disabled | Disables computing feature importance within a signal. |
| Enabled | Enabled | Enables computing feature importance within a signal. |
| Name | Value | Description |
|---|
| FeatureImportanceModeDisabled | Disabled | Disables computing feature importance within a signal. |
| FeatureImportanceModeEnabled | Enabled | Enables computing feature importance within a signal. |
| Value | Description |
|---|
Disabled | Disables computing feature importance within a signal. |
Enabled | Enables computing feature importance within a signal. |
| Name | Value | Description |
|---|
| Disabled | Disabled | Disables computing feature importance within a signal. |
| Enabled | Enabled | Enables computing feature importance within a signal. |
| Name | Value | Description |
|---|
| Disabled | Disabled | Disables computing feature importance within a signal. |
| Enabled | Enabled | Enables computing feature importance within a signal. |
| Name | Value | Description |
|---|
| DISABLED | Disabled | Disables computing feature importance within a signal. |
| ENABLED | Enabled | Enables computing feature importance within a signal. |
| Value | Description |
|---|
Disabled | Disables computing feature importance within a signal. |
Enabled | Enables computing feature importance within a signal. |
FeatureImportanceSettings, FeatureImportanceSettingsArgs
FeatureImportanceSettingsResponse
- Mode
string
- The mode of operation for computing feature importance.
- TargetColumn
string
- The name of the target column within the input data asset.
- Mode
string
- The mode of operation for computing feature importance.
- TargetColumn
string
- The name of the target column within the input data asset.
- mode
string
- The mode of operation for computing feature importance.
- target_column
string
- The name of the target column within the input data asset.
- mode
String
- The mode of operation for computing feature importance.
- targetColumn
String
- The name of the target column within the input data asset.
- mode
string
- The mode of operation for computing feature importance.
- targetColumn
string
- The name of the target column within the input data asset.
- mode
str
- The mode of operation for computing feature importance.
- target_column
str
- The name of the target column within the input data asset.
- mode
String
- 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 | Value | Description |
|---|
| None | None | No feature lags generated. |
| Auto | Auto | System auto-generates feature lags. |
| Name | Value | Description |
|---|
| FeatureLagsNone | None | No feature lags generated. |
| FeatureLagsAuto | Auto | System auto-generates feature lags. |
| Value | Description |
|---|
None | No feature lags generated. |
Auto | System auto-generates feature lags. |
| Name | Value | Description |
|---|
| None | None | No feature lags generated. |
| Auto | Auto | System auto-generates feature lags. |
| Name | Value | Description |
|---|
| None | None | No feature lags generated. |
| Auto | Auto | System auto-generates feature lags. |
| Name | Value | Description |
|---|
| NONE | None | No feature lags generated. |
| AUTO | Auto | System auto-generates feature lags. |
| Value | Description |
|---|
None | No feature lags generated. |
Auto | System auto-generates feature lags. |
FeatureSubset, FeatureSubsetArgs
- Features
List<string>
- [Required] The list of features to include.
- Features
[]string
- [Required] The list of features to include.
- features
list(string)
- [Required] The list of features to include.
- features
List<String>
- [Required] The list of features to include.
- features
string[]
- [Required] The list of features to include.
- features
Sequence[str]
- [Required] The list of features to include.
- features
List<String>
- [Required] The list of features to include.
FeatureSubsetResponse
- Features
List<string>
- [Required] The list of features to include.
- Features
[]string
- [Required] The list of features to include.
- features
list(string)
- [Required] The list of features to include.
- features
List<String>
- [Required] The list of features to include.
- features
string[]
- [Required] The list of features to include.
- features
Sequence[str]
- [Required] The list of features to include.
- features
List<String>
- [Required] The list of features to include.
FeaturizationMode
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.
| Name | Value | Description |
|---|
| Auto | Auto | Auto mode, system performs featurization without any custom featurization inputs. |
| Custom | Custom | Custom featurization. |
| Off | Off | Featurization off. 'Forecasting' task cannot use this value. |
| Name | Value | Description |
|---|
| FeaturizationModeAuto | Auto | Auto mode, system performs featurization without any custom featurization inputs. |
| FeaturizationModeCustom | Custom | Custom featurization. |
| FeaturizationModeOff | Off | Featurization off. 'Forecasting' task cannot use this value. |
| Value | Description |
|---|
Auto | Auto mode, system performs featurization without any custom featurization inputs. |
Custom | Custom featurization. |
Off | Featurization off. 'Forecasting' task cannot use this value. |
| Name | Value | Description |
|---|
| Auto | Auto | Auto mode, system performs featurization without any custom featurization inputs. |
| Custom | Custom | Custom featurization. |
| Off | Off | Featurization off. 'Forecasting' task cannot use this value. |
| Name | Value | Description |
|---|
| Auto | Auto | Auto mode, system performs featurization without any custom featurization inputs. |
| Custom | Custom | Custom featurization. |
| Off | Off | Featurization off. 'Forecasting' task cannot use this value. |
| Name | Value | Description |
|---|
| AUTO | Auto | Auto mode, system performs featurization without any custom featurization inputs. |
| CUSTOM | Custom | Custom featurization. |
| OFF | Off | Featurization off. 'Forecasting' task cannot use this value. |
| Value | Description |
|---|
Auto | Auto mode, system performs featurization without any custom featurization inputs. |
Custom | Custom featurization. |
Off | Featurization off. 'Forecasting' task cannot use this value. |
Fixed input data definition.
- JobInputType
string | JobInputType
- [Required] Specifies the type of job.
- Uri
string
- [Required] Input Asset URI.
- Columns
map[string]string
- Mapping of column names to special uses.
- DataContext
string
- The context metadata of the data source.
- jobInputType
String | JobInputType
- [Required] Specifies the type of job.
- uri
String
- [Required] Input Asset URI.
- columns
Map<String,String>
- Mapping of column names to special uses.
- dataContext
String
- The context metadata of the data source.
- jobInputType
string | JobInputType
- [Required] Specifies the type of job.
- uri
string
- [Required] Input Asset URI.
- columns
{[key: string]: string}
- Mapping of column names to special uses.
- dataContext
string
- The context metadata of the data source.
Fixed input data definition.
- JobInputType
string
- [Required] Specifies the type of job.
- Uri
string
- [Required] Input Asset URI.
- Columns
Dictionary<string, string>
- Mapping of column names to special uses.
- DataContext
string
- The context metadata of the data source.
- JobInputType
string
- [Required] Specifies the type of job.
- Uri
string
- [Required] Input Asset URI.
- Columns
map[string]string
- Mapping of column names to special uses.
- DataContext
string
- The context metadata of the data source.
- job_input_type
string
- [Required] Specifies the type of job.
- uri
string
- [Required] Input Asset URI.
- columns
map(string)
- Mapping of column names to special uses.
- data_context
string
- The context metadata of the data source.
- jobInputType
String
- [Required] Specifies the type of job.
- uri
String
- [Required] Input Asset URI.
- columns
Map<String,String>
- Mapping of column names to special uses.
- dataContext
String
- The context metadata of the data source.
- jobInputType
string
- [Required] Specifies the type of job.
- uri
string
- [Required] Input Asset URI.
- columns
{[key: string]: string}
- Mapping of column names to special uses.
- dataContext
string
- The context metadata of the data source.
- job_input_type
str
- [Required] Specifies the type of job.
- uri
str
- [Required] Input Asset URI.
- columns
Mapping[str, str]
- Mapping of column names to special uses.
- data_context
str
- The context metadata of the data source.
- jobInputType
String
- [Required] Specifies the type of job.
- uri
String
- [Required] Input Asset URI.
- columns
Map<String>
- Mapping of column names to special uses.
- dataContext
String
- The context metadata of the data source.
Forecasting, ForecastingArgs
Forecasting task in AutoML Table vertical.
ForecastingModels
Enum for all forecasting models supported by AutoML.
| Name | Value | Description |
|---|
| AutoArima | AutoArima | 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. |
| Prophet | Prophet | 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. |
| Naive | Naive | The Naive forecasting model makes predictions by carrying forward the latest target value for each time-series in the training data. |
| SeasonalNaive | SeasonalNaive | The Seasonal Naive forecasting model makes predictions by carrying forward the latest season of target values for each time-series in the training data. |
| Average | Average | The Average forecasting model makes predictions by carrying forward the average of the target values for each time-series in the training data. |
| SeasonalAverage | SeasonalAverage | 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. |
| ExponentialSmoothing | ExponentialSmoothing | 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. |
| Arimax | Arimax | 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. |
| TCNForecaster | TCNForecaster | TCNForecaster: Temporal Convolutional Networks Forecaster. //TODO: Ask forecasting team for brief intro. |
| ElasticNet | ElasticNet | Elastic net is a popular type of regularized linear regression that combines two popular penalties, specifically the L1 and L2 penalty functions. |
| GradientBoosting | GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
| DecisionTree | DecisionTree | 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. |
| KNN | KNN | 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 | LassoLars | Lasso model fit with Least Angle Regression a.k.a. Lars. It is a Linear Model trained with an L1 prior as regularizer. |
| SGD | SGD | 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. |
| RandomForest | RandomForest | 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. |
| ExtremeRandomTrees | ExtremeRandomTrees | 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. |
| LightGBM | LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
| XGBoostRegressor | XGBoostRegressor | XGBoostRegressor: Extreme Gradient Boosting Regressor is a supervised machine learning model using ensemble of base learners. |
| Name | Value | Description |
|---|
| ForecastingModelsAutoArima | AutoArima | 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. |
| ForecastingModelsProphet | Prophet | 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. |
| ForecastingModelsNaive | Naive | The Naive forecasting model makes predictions by carrying forward the latest target value for each time-series in the training data. |
| ForecastingModelsSeasonalNaive | SeasonalNaive | The Seasonal Naive forecasting model makes predictions by carrying forward the latest season of target values for each time-series in the training data. |
| ForecastingModelsAverage | Average | The Average forecasting model makes predictions by carrying forward the average of the target values for each time-series in the training data. |
| ForecastingModelsSeasonalAverage | SeasonalAverage | 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. |
| ForecastingModelsExponentialSmoothing | ExponentialSmoothing | 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. |
| ForecastingModelsArimax | Arimax | 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. |
| ForecastingModelsTCNForecaster | TCNForecaster | TCNForecaster: Temporal Convolutional Networks Forecaster. //TODO: Ask forecasting team for brief intro. |
| ForecastingModelsElasticNet | ElasticNet | Elastic net is a popular type of regularized linear regression that combines two popular penalties, specifically the L1 and L2 penalty functions. |
| ForecastingModelsGradientBoosting | GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
| ForecastingModelsDecisionTree | DecisionTree | 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. |
| ForecastingModelsKNN | KNN | 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. |
| ForecastingModelsLassoLars | LassoLars | Lasso model fit with Least Angle Regression a.k.a. Lars. It is a Linear Model trained with an L1 prior as regularizer. |
| ForecastingModelsSGD | SGD | 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. |
| ForecastingModelsRandomForest | RandomForest | 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. |
| ForecastingModelsExtremeRandomTrees | ExtremeRandomTrees | 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. |
| ForecastingModelsLightGBM | LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
| ForecastingModelsXGBoostRegressor | XGBoostRegressor | XGBoostRegressor: Extreme Gradient Boosting Regressor is a supervised machine learning model using ensemble of base learners. |
| Value | Description |
|---|
AutoArima | 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. |
Prophet | 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. |
Naive | The Naive forecasting model makes predictions by carrying forward the latest target value for each time-series in the training data. |
SeasonalNaive | The Seasonal Naive forecasting model makes predictions by carrying forward the latest season of target values for each time-series in the training data. |
Average | The Average forecasting model makes predictions by carrying forward the average of the target values for each time-series in the training data. |
SeasonalAverage | 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. |
ExponentialSmoothing | 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. |
Arimax | 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. |
TCNForecaster | TCNForecaster: Temporal Convolutional Networks Forecaster. //TODO: Ask forecasting team for brief intro. |
ElasticNet | Elastic net is a popular type of regularized linear regression that combines two popular penalties, specifically the L1 and L2 penalty functions. |
GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
DecisionTree | 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. |
KNN | 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 | Lasso model fit with Least Angle Regression a.k.a. Lars. It is a Linear Model trained with an L1 prior as regularizer. |
SGD | 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. |
RandomForest | 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. |
ExtremeRandomTrees | 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. |
LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
XGBoostRegressor | XGBoostRegressor: Extreme Gradient Boosting Regressor is a supervised machine learning model using ensemble of base learners. |
| Name | Value | Description |
|---|
| AutoArima | AutoArima | 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. |
| Prophet | Prophet | 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. |
| Naive | Naive | The Naive forecasting model makes predictions by carrying forward the latest target value for each time-series in the training data. |
| SeasonalNaive | SeasonalNaive | The Seasonal Naive forecasting model makes predictions by carrying forward the latest season of target values for each time-series in the training data. |
| Average | Average | The Average forecasting model makes predictions by carrying forward the average of the target values for each time-series in the training data. |
| SeasonalAverage | SeasonalAverage | 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. |
| ExponentialSmoothing | ExponentialSmoothing | 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. |
| Arimax | Arimax | 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. |
| TCNForecaster | TCNForecaster | TCNForecaster: Temporal Convolutional Networks Forecaster. //TODO: Ask forecasting team for brief intro. |
| ElasticNet | ElasticNet | Elastic net is a popular type of regularized linear regression that combines two popular penalties, specifically the L1 and L2 penalty functions. |
| GradientBoosting | GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
| DecisionTree | DecisionTree | 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. |
| KNN | KNN | 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 | LassoLars | Lasso model fit with Least Angle Regression a.k.a. Lars. It is a Linear Model trained with an L1 prior as regularizer. |
| SGD | SGD | 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. |
| RandomForest | RandomForest | 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. |
| ExtremeRandomTrees | ExtremeRandomTrees | 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. |
| LightGBM | LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
| XGBoostRegressor | XGBoostRegressor | XGBoostRegressor: Extreme Gradient Boosting Regressor is a supervised machine learning model using ensemble of base learners. |
| Name | Value | Description |
|---|
| AutoArima | AutoArima | 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. |
| Prophet | Prophet | 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. |
| Naive | Naive | The Naive forecasting model makes predictions by carrying forward the latest target value for each time-series in the training data. |
| SeasonalNaive | SeasonalNaive | The Seasonal Naive forecasting model makes predictions by carrying forward the latest season of target values for each time-series in the training data. |
| Average | Average | The Average forecasting model makes predictions by carrying forward the average of the target values for each time-series in the training data. |
| SeasonalAverage | SeasonalAverage | 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. |
| ExponentialSmoothing | ExponentialSmoothing | 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. |
| Arimax | Arimax | 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. |
| TCNForecaster | TCNForecaster | TCNForecaster: Temporal Convolutional Networks Forecaster. //TODO: Ask forecasting team for brief intro. |
| ElasticNet | ElasticNet | Elastic net is a popular type of regularized linear regression that combines two popular penalties, specifically the L1 and L2 penalty functions. |
| GradientBoosting | GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
| DecisionTree | DecisionTree | 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. |
| KNN | KNN | 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 | LassoLars | Lasso model fit with Least Angle Regression a.k.a. Lars. It is a Linear Model trained with an L1 prior as regularizer. |
| SGD | SGD | 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. |
| RandomForest | RandomForest | 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. |
| ExtremeRandomTrees | ExtremeRandomTrees | 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. |
| LightGBM | LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
| XGBoostRegressor | XGBoostRegressor | XGBoostRegressor: Extreme Gradient Boosting Regressor is a supervised machine learning model using ensemble of base learners. |
| Name | Value | Description |
|---|
| AUTO_ARIMA | AutoArima | 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. |
| PROPHET | Prophet | 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. |
| NAIVE | Naive | The Naive forecasting model makes predictions by carrying forward the latest target value for each time-series in the training data. |
| SEASONAL_NAIVE | SeasonalNaive | The Seasonal Naive forecasting model makes predictions by carrying forward the latest season of target values for each time-series in the training data. |
| AVERAGE | Average | The Average forecasting model makes predictions by carrying forward the average of the target values for each time-series in the training data. |
| SEASONAL_AVERAGE | SeasonalAverage | 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. |
| EXPONENTIAL_SMOOTHING | ExponentialSmoothing | 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. |
| ARIMAX | Arimax | 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. |
| TCN_FORECASTER | TCNForecaster | TCNForecaster: Temporal Convolutional Networks Forecaster. //TODO: Ask forecasting team for brief intro. |
| ELASTIC_NET | ElasticNet | Elastic net is a popular type of regularized linear regression that combines two popular penalties, specifically the L1 and L2 penalty functions. |
| GRADIENT_BOOSTING | GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
| DECISION_TREE | DecisionTree | 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. |
| KNN | KNN | 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. |
| LASSO_LARS | LassoLars | Lasso model fit with Least Angle Regression a.k.a. Lars. It is a Linear Model trained with an L1 prior as regularizer. |
| SGD | SGD | 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. |
| RANDOM_FOREST | RandomForest | 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. |
| EXTREME_RANDOM_TREES | ExtremeRandomTrees | 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. |
| LIGHT_GBM | LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
| XG_BOOST_REGRESSOR | XGBoostRegressor | XGBoostRegressor: Extreme Gradient Boosting Regressor is a supervised machine learning model using ensemble of base learners. |
| Value | Description |
|---|
AutoArima | 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. |
Prophet | 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. |
Naive | The Naive forecasting model makes predictions by carrying forward the latest target value for each time-series in the training data. |
SeasonalNaive | The Seasonal Naive forecasting model makes predictions by carrying forward the latest season of target values for each time-series in the training data. |
Average | The Average forecasting model makes predictions by carrying forward the average of the target values for each time-series in the training data. |
SeasonalAverage | 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. |
ExponentialSmoothing | 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. |
Arimax | 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. |
TCNForecaster | TCNForecaster: Temporal Convolutional Networks Forecaster. //TODO: Ask forecasting team for brief intro. |
ElasticNet | Elastic net is a popular type of regularized linear regression that combines two popular penalties, specifically the L1 and L2 penalty functions. |
GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
DecisionTree | 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. |
KNN | 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 | Lasso model fit with Least Angle Regression a.k.a. Lars. It is a Linear Model trained with an L1 prior as regularizer. |
SGD | 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. |
RandomForest | 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. |
ExtremeRandomTrees | 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. |
LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
XGBoostRegressor | XGBoostRegressor: Extreme Gradient Boosting Regressor is a supervised machine learning model using ensemble of base learners. |
ForecastingPrimaryMetrics
Primary metrics for Forecasting task.
| Name | Value | Description |
|---|
| SpearmanCorrelation | SpearmanCorrelation | The Spearman's rank coefficient of correlation is a non-parametric measure of rank correlation. |
| NormalizedRootMeanSquaredError | NormalizedRootMeanSquaredError | The Normalized Root Mean Squared Error (NRMSE) the RMSE facilitates the comparison between models with different scales. |
| R2Score | R2Score | The R2 score is one of the performance evaluation measures for forecasting-based machine learning models. |
| NormalizedMeanAbsoluteError | NormalizedMeanAbsoluteError | The Normalized Mean Absolute Error (NMAE) is a validation metric to compare the Mean Absolute Error (MAE) of (time) series with different scales. |
| Name | Value | Description |
|---|
| ForecastingPrimaryMetricsSpearmanCorrelation | SpearmanCorrelation | The Spearman's rank coefficient of correlation is a non-parametric measure of rank correlation. |
| ForecastingPrimaryMetricsNormalizedRootMeanSquaredError | NormalizedRootMeanSquaredError | The Normalized Root Mean Squared Error (NRMSE) the RMSE facilitates the comparison between models with different scales. |
| ForecastingPrimaryMetricsR2Score | R2Score | The R2 score is one of the performance evaluation measures for forecasting-based machine learning models. |
| ForecastingPrimaryMetricsNormalizedMeanAbsoluteError | NormalizedMeanAbsoluteError | The Normalized Mean Absolute Error (NMAE) is a validation metric to compare the Mean Absolute Error (MAE) of (time) series with different scales. |
| Value | Description |
|---|
SpearmanCorrelation | The Spearman's rank coefficient of correlation is a non-parametric measure of rank correlation. |
NormalizedRootMeanSquaredError | The Normalized Root Mean Squared Error (NRMSE) the RMSE facilitates the comparison between models with different scales. |
R2Score | The R2 score is one of the performance evaluation measures for forecasting-based machine learning models. |
NormalizedMeanAbsoluteError | The Normalized Mean Absolute Error (NMAE) is a validation metric to compare the Mean Absolute Error (MAE) of (time) series with different scales. |
| Name | Value | Description |
|---|
| SpearmanCorrelation | SpearmanCorrelation | The Spearman's rank coefficient of correlation is a non-parametric measure of rank correlation. |
| NormalizedRootMeanSquaredError | NormalizedRootMeanSquaredError | The Normalized Root Mean Squared Error (NRMSE) the RMSE facilitates the comparison between models with different scales. |
| R2Score | R2Score | The R2 score is one of the performance evaluation measures for forecasting-based machine learning models. |
| NormalizedMeanAbsoluteError | NormalizedMeanAbsoluteError | The Normalized Mean Absolute Error (NMAE) is a validation metric to compare the Mean Absolute Error (MAE) of (time) series with different scales. |
| Name | Value | Description |
|---|
| SpearmanCorrelation | SpearmanCorrelation | The Spearman's rank coefficient of correlation is a non-parametric measure of rank correlation. |
| NormalizedRootMeanSquaredError | NormalizedRootMeanSquaredError | The Normalized Root Mean Squared Error (NRMSE) the RMSE facilitates the comparison between models with different scales. |
| R2Score | R2Score | The R2 score is one of the performance evaluation measures for forecasting-based machine learning models. |
| NormalizedMeanAbsoluteError | NormalizedMeanAbsoluteError | The Normalized Mean Absolute Error (NMAE) is a validation metric to compare the Mean Absolute Error (MAE) of (time) series with different scales. |
| Name | Value | Description |
|---|
| SPEARMAN_CORRELATION | SpearmanCorrelation | The Spearman's rank coefficient of correlation is a non-parametric measure of rank correlation. |
| NORMALIZED_ROOT_MEAN_SQUARED_ERROR | NormalizedRootMeanSquaredError | The Normalized Root Mean Squared Error (NRMSE) the RMSE facilitates the comparison between models with different scales. |
| R2_SCORE | R2Score | The R2 score is one of the performance evaluation measures for forecasting-based machine learning models. |
| NORMALIZED_MEAN_ABSOLUTE_ERROR | NormalizedMeanAbsoluteError | The Normalized Mean Absolute Error (NMAE) is a validation metric to compare the Mean Absolute Error (MAE) of (time) series with different scales. |
| Value | Description |
|---|
SpearmanCorrelation | The Spearman's rank coefficient of correlation is a non-parametric measure of rank correlation. |
NormalizedRootMeanSquaredError | The Normalized Root Mean Squared Error (NRMSE) the RMSE facilitates the comparison between models with different scales. |
R2Score | The R2 score is one of the performance evaluation measures for forecasting-based machine learning models. |
NormalizedMeanAbsoluteError | The Normalized Mean Absolute Error (NMAE) is a validation metric to compare the Mean Absolute Error (MAE) of (time) series with different scales. |
ForecastingResponse
Forecasting task in AutoML Table vertical.
ForecastingSettings, ForecastingSettingsArgs
Forecasting specific parameters.
- 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
int
- Number of periods between the origin time of one CV fold and the next fold. For
example, if
CVStepSize = 3 for daily data, the origin time for each fold will be
three days apart. - FeatureLags
string | Pulumi.AzureNative.MachineLearningServices.FeatureLags
- Flag for generating lags for the numeric features.
- ForecastHorizon
Pulumi.AzureNative.MachineLearningServices.Inputs.AutoForecastHorizon | Pulumi.AzureNative.MachineLearningServices.Inputs.CustomForecastHorizon
- 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
Pulumi.AzureNative.MachineLearningServices.Inputs.AutoSeasonality | Pulumi.AzureNative.MachineLearningServices.Inputs.CustomSeasonality
- Set time series seasonality as an integer multiple of the series frequency.
If seasonality is set to 'auto', it will be inferred.
- ShortSeriesHandlingConfig
string | Pulumi.AzureNative.MachineLearningServices.ShortSeriesHandlingConfiguration
- The parameter defining how if AutoML should handle short time series.
- TargetAggregateFunction
string | Pulumi.AzureNative.MachineLearningServices.TargetAggregationFunction
- Target aggregate function.
- TargetLags
Pulumi.AzureNative.MachineLearningServices.Inputs.AutoTargetLags | Pulumi.AzureNative.MachineLearningServices.Inputs.CustomTargetLags
- The number of past periods to lag from the target column.
- TargetRollingWindowSize
Pulumi.AzureNative.MachineLearningServices.Inputs.AutoTargetRollingWindowSize | Pulumi.AzureNative.MachineLearningServices.Inputs.CustomTargetRollingWindowSize
- 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
List<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
string | Pulumi.AzureNative.MachineLearningServices.UseStl
- Configure STL Decomposition of the time-series target column.
- 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
int
- Number of periods between the origin time of one CV fold and the next fold. For
example, if
CVStepSize = 3 for daily data, the origin time for each fold will be
three days apart. - FeatureLags
string | FeatureLags
- Flag for generating lags for the numeric features.
- ForecastHorizon
AutoForecastHorizon | CustomForecastHorizon
- 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
AutoSeasonality | CustomSeasonality
- Set time series seasonality as an integer multiple of the series frequency.
If seasonality is set to 'auto', it will be inferred.
- ShortSeriesHandlingConfig
string | ShortSeriesHandlingConfiguration
- The parameter defining how if AutoML should handle short time series.
- TargetAggregateFunction
string | TargetAggregationFunction
- Target aggregate function.
- TargetLags
AutoTargetLags | CustomTargetLags
- The number of past periods to lag from the target column.
- TargetRollingWindowSize
AutoTargetRollingWindowSize | CustomTargetRollingWindowSize
- 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
string | UseStl
- Configure STL Decomposition of the time-series target column.
- country_or_region_for_holidays
string
- Country or region for holidays for forecasting tasks.
These should be ISO 3166 two-letter country/region codes, for example 'US' or 'GB'.
- cv_step_size
number
- Number of periods between the origin time of one CV fold and the next fold. For
example, if
CVStepSize = 3 for daily data, the origin time for each fold will be
three days apart. - feature_lags
string | "None" | "Auto"
- Flag for generating lags for the numeric features.
- forecast_horizon
object | object
- 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
object | object
- Set time series seasonality as an integer multiple of the series frequency.
If seasonality is set to 'auto', it will be inferred.
- short_series_handling_config
string | "None" | "Auto" | "Pad" | "Drop"
- The parameter defining how if AutoML should handle short time series.
- target_aggregate_function
string | "None" | "Sum" | "Max" | "Min" | "Mean"
- Target aggregate function.
- target_lags
object | object
- The number of past periods to lag from the target column.
- target_rolling_window_size
object | object
- The number of past periods used to create a rolling window average of the target column.
- time_column_name
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.
- time_series_id_column_names
list(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.
- use_stl
string | "None" | "Season" | "SeasonTrend"
- Configure STL Decomposition of the time-series target column.
- 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
CVStepSize = 3 for daily data, the origin time for each fold will be
three days apart. - featureLags
String | FeatureLags
- Flag for generating lags for the numeric features.
- forecastHorizon
AutoForecastHorizon | CustomForecastHorizon
- 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
AutoSeasonality | CustomSeasonality
- Set time series seasonality as an integer multiple of the series frequency.
If seasonality is set to 'auto', it will be inferred.
- shortSeriesHandlingConfig
String | ShortSeriesHandlingConfiguration
- The parameter defining how if AutoML should handle short time series.
- targetAggregateFunction
String | TargetAggregationFunction
- Target aggregate function.
- targetLags
AutoTargetLags | CustomTargetLags
- The number of past periods to lag from the target column.
- targetRollingWindowSize
AutoTargetRollingWindowSize | CustomTargetRollingWindowSize
- 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
List<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
String | UseStl
- Configure STL Decomposition of the time-series target column.
- 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
number
- Number of periods between the origin time of one CV fold and the next fold. For
example, if
CVStepSize = 3 for daily data, the origin time for each fold will be
three days apart. - featureLags
string | FeatureLags
- Flag for generating lags for the numeric features.
- forecastHorizon
AutoForecastHorizon | CustomForecastHorizon
- 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
AutoSeasonality | CustomSeasonality
- Set time series seasonality as an integer multiple of the series frequency.
If seasonality is set to 'auto', it will be inferred.
- shortSeriesHandlingConfig
string | ShortSeriesHandlingConfiguration
- The parameter defining how if AutoML should handle short time series.
- targetAggregateFunction
string | TargetAggregationFunction
- Target aggregate function.
- targetLags
AutoTargetLags | CustomTargetLags
- The number of past periods to lag from the target column.
- targetRollingWindowSize
AutoTargetRollingWindowSize | CustomTargetRollingWindowSize
- 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
string | UseStl
- Configure STL Decomposition of the time-series target column.
- country_or_region_for_holidays
str
- Country or region for holidays for forecasting tasks.
These should be ISO 3166 two-letter country/region codes, for example 'US' or 'GB'.
- cv_step_size
int
- Number of periods between the origin time of one CV fold and the next fold. For
example, if
CVStepSize = 3 for daily data, the origin time for each fold will be
three days apart. - feature_lags
str | FeatureLags
- Flag for generating lags for the numeric features.
- forecast_horizon
AutoForecastHorizon | CustomForecastHorizon
- The desired maximum forecast horizon in units of time-series frequency.
- frequency
str
- 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
AutoSeasonality | CustomSeasonality
- Set time series seasonality as an integer multiple of the series frequency.
If seasonality is set to 'auto', it will be inferred.
- short_series_handling_config
str | ShortSeriesHandlingConfiguration
- The parameter defining how if AutoML should handle short time series.
- target_aggregate_function
str | TargetAggregationFunction
- Target aggregate function.
- target_lags
AutoTargetLags | CustomTargetLags
- The number of past periods to lag from the target column.
- target_rolling_window_size
AutoTargetRollingWindowSize | CustomTargetRollingWindowSize
- The number of past periods used to create a rolling window average of the target column.
- time_column_name
str
- 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.
- time_series_id_column_names
Sequence[str]
- 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.
- use_stl
str | UseStl
- Configure STL Decomposition of the time-series target column.
- 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
Number
- Number of periods between the origin time of one CV fold and the next fold. For
example, if
CVStepSize = 3 for daily data, the origin time for each fold will be
three days apart. - featureLags
String | "None" | "Auto"
- Flag for generating lags for the numeric features.
- forecastHorizon
Property Map | Property Map
- 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
Property Map | Property Map
- Set time series seasonality as an integer multiple of the series frequency.
If seasonality is set to 'auto', it will be inferred.
- shortSeriesHandlingConfig
String | "None" | "Auto" | "Pad" | "Drop"
- The parameter defining how if AutoML should handle short time series.
- targetAggregateFunction
String | "None" | "Sum" | "Max" | "Min" | "Mean"
- Target aggregate function.
- targetLags
Property Map | Property Map
- The number of past periods to lag from the target column.
- targetRollingWindowSize
Property Map | Property Map
- 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
List<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
String | "None" | "Season" | "SeasonTrend"
- Configure STL Decomposition of the time-series target column.
ForecastingSettingsResponse
Forecasting specific parameters.
- 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
int
- Number of periods between the origin time of one CV fold and the next fold. For
example, if
CVStepSize = 3 for daily data, the origin time for each fold will be
three days apart. - FeatureLags
string
- Flag for generating lags for the numeric features.
- ForecastHorizon
Pulumi.AzureNative.MachineLearningServices.Inputs.AutoForecastHorizonResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.CustomForecastHorizonResponse
- 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
Pulumi.AzureNative.MachineLearningServices.Inputs.AutoSeasonalityResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.CustomSeasonalityResponse
- Set time series seasonality as an integer multiple of the series frequency.
If seasonality is set to 'auto', it will be inferred.
- ShortSeriesHandlingConfig
string
- The parameter defining how if AutoML should handle short time series.
- TargetAggregateFunction
string
- Target aggregate function.
- TargetLags
Pulumi.AzureNative.MachineLearningServices.Inputs.AutoTargetLagsResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.CustomTargetLagsResponse
- The number of past periods to lag from the target column.
- TargetRollingWindowSize
Pulumi.AzureNative.MachineLearningServices.Inputs.AutoTargetRollingWindowSizeResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.CustomTargetRollingWindowSizeResponse
- 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
List<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
string
- Configure STL Decomposition of the time-series target column.
- 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
int
- Number of periods between the origin time of one CV fold and the next fold. For
example, if
CVStepSize = 3 for daily data, the origin time for each fold will be
three days apart. - FeatureLags
string
- Flag for generating lags for the numeric features.
- ForecastHorizon
AutoForecastHorizonResponse | CustomForecastHorizonResponse
- 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
AutoSeasonalityResponse | CustomSeasonalityResponse
- Set time series seasonality as an integer multiple of the series frequency.
If seasonality is set to 'auto', it will be inferred.
- ShortSeriesHandlingConfig
string
- The parameter defining how if AutoML should handle short time series.
- TargetAggregateFunction
string
- Target aggregate function.
- TargetLags
AutoTargetLagsResponse | CustomTargetLagsResponse
- The number of past periods to lag from the target column.
- TargetRollingWindowSize
AutoTargetRollingWindowSizeResponse | CustomTargetRollingWindowSizeResponse
- 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
string
- Configure STL Decomposition of the time-series target column.
- country_or_region_for_holidays
string
- Country or region for holidays for forecasting tasks.
These should be ISO 3166 two-letter country/region codes, for example 'US' or 'GB'.
- cv_step_size
number
- Number of periods between the origin time of one CV fold and the next fold. For
example, if
CVStepSize = 3 for daily data, the origin time for each fold will be
three days apart. - feature_lags
string
- Flag for generating lags for the numeric features.
- forecast_horizon
object | object
- 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
object | object
- Set time series seasonality as an integer multiple of the series frequency.
If seasonality is set to 'auto', it will be inferred.
- short_series_handling_config
string
- The parameter defining how if AutoML should handle short time series.
- target_aggregate_function
string
- Target aggregate function.
- target_lags
object | object
- The number of past periods to lag from the target column.
- target_rolling_window_size
object | object
- The number of past periods used to create a rolling window average of the target column.
- time_column_name
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.
- time_series_id_column_names
list(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.
- use_stl
string
- Configure STL Decomposition of the time-series target column.
- 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
CVStepSize = 3 for daily data, the origin time for each fold will be
three days apart. - featureLags
String
- Flag for generating lags for the numeric features.
- forecastHorizon
AutoForecastHorizonResponse | CustomForecastHorizonResponse
- 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
AutoSeasonalityResponse | CustomSeasonalityResponse
- Set time series seasonality as an integer multiple of the series frequency.
If seasonality is set to 'auto', it will be inferred.
- shortSeriesHandlingConfig
String
- The parameter defining how if AutoML should handle short time series.
- targetAggregateFunction
String
- Target aggregate function.
- targetLags
AutoTargetLagsResponse | CustomTargetLagsResponse
- The number of past periods to lag from the target column.
- targetRollingWindowSize
AutoTargetRollingWindowSizeResponse | CustomTargetRollingWindowSizeResponse
- 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
List<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
String
- Configure STL Decomposition of the time-series target column.
- 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
number
- Number of periods between the origin time of one CV fold and the next fold. For
example, if
CVStepSize = 3 for daily data, the origin time for each fold will be
three days apart. - featureLags
string
- Flag for generating lags for the numeric features.
- forecastHorizon
AutoForecastHorizonResponse | CustomForecastHorizonResponse
- 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
AutoSeasonalityResponse | CustomSeasonalityResponse
- Set time series seasonality as an integer multiple of the series frequency.
If seasonality is set to 'auto', it will be inferred.
- shortSeriesHandlingConfig
string
- The parameter defining how if AutoML should handle short time series.
- targetAggregateFunction
string
- Target aggregate function.
- targetLags
AutoTargetLagsResponse | CustomTargetLagsResponse
- The number of past periods to lag from the target column.
- targetRollingWindowSize
AutoTargetRollingWindowSizeResponse | CustomTargetRollingWindowSizeResponse
- 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
string
- Configure STL Decomposition of the time-series target column.
- country_or_region_for_holidays
str
- Country or region for holidays for forecasting tasks.
These should be ISO 3166 two-letter country/region codes, for example 'US' or 'GB'.
- cv_step_size
int
- Number of periods between the origin time of one CV fold and the next fold. For
example, if
CVStepSize = 3 for daily data, the origin time for each fold will be
three days apart. - feature_lags
str
- Flag for generating lags for the numeric features.
- forecast_horizon
AutoForecastHorizonResponse | CustomForecastHorizonResponse
- The desired maximum forecast horizon in units of time-series frequency.
- frequency
str
- 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
AutoSeasonalityResponse | CustomSeasonalityResponse
- Set time series seasonality as an integer multiple of the series frequency.
If seasonality is set to 'auto', it will be inferred.
- short_series_handling_config
str
- The parameter defining how if AutoML should handle short time series.
- target_aggregate_function
str
- Target aggregate function.
- target_lags
AutoTargetLagsResponse | CustomTargetLagsResponse
- The number of past periods to lag from the target column.
- target_rolling_window_size
AutoTargetRollingWindowSizeResponse | CustomTargetRollingWindowSizeResponse
- The number of past periods used to create a rolling window average of the target column.
- time_column_name
str
- 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.
- time_series_id_column_names
Sequence[str]
- 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.
- use_stl
str
- Configure STL Decomposition of the time-series target column.
- 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
Number
- Number of periods between the origin time of one CV fold and the next fold. For
example, if
CVStepSize = 3 for daily data, the origin time for each fold will be
three days apart. - featureLags
String
- Flag for generating lags for the numeric features.
- forecastHorizon
Property Map | Property Map
- 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
Property Map | Property Map
- Set time series seasonality as an integer multiple of the series frequency.
If seasonality is set to 'auto', it will be inferred.
- shortSeriesHandlingConfig
String
- The parameter defining how if AutoML should handle short time series.
- targetAggregateFunction
String
- Target aggregate function.
- targetLags
Property Map | Property Map
- The number of past periods to lag from the target column.
- targetRollingWindowSize
Property Map | Property Map
- 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
List<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
String
- Configure STL Decomposition of the time-series target column.
ForecastingTrainingSettings, ForecastingTrainingSettingsArgs
Forecasting Training related configuration.
- allowed_training_algorithms
list(string | "AutoArima" | "Prophet" | "Naive" | "SeasonalNaive" | "Average" | "SeasonalAverage" | "ExponentialSmoothing" | "Arimax" | "TCNForecaster" | "ElasticNet" | "GradientBoosting" | "DecisionTree" | "KNN" | "LassoLars" | "SGD" | "RandomForest" | "ExtremeRandomTrees" | "LightGBM" | "XGBoostRegressor")
- Allowed models for forecasting task.
- blocked_training_algorithms
list(string | "AutoArima" | "Prophet" | "Naive" | "SeasonalNaive" | "Average" | "SeasonalAverage" | "ExponentialSmoothing" | "Arimax" | "TCNForecaster" | "ElasticNet" | "GradientBoosting" | "DecisionTree" | "KNN" | "LassoLars" | "SGD" | "RandomForest" | "ExtremeRandomTrees" | "LightGBM" | "XGBoostRegressor")
- Blocked models for forecasting task.
- enable_dnn_training
bool
- Enable recommendation of DNN models.
- enable_model_explainability
bool
- Flag to turn on explainability on best model.
- enable_onnx_compatible_models
bool
- Flag for enabling onnx compatible models.
- enable_stack_ensemble
bool
- Enable stack ensemble run.
- enable_vote_ensemble
bool
- Enable voting ensemble run.
- ensemble_model_download_timeout
string
- 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.
- stack_ensemble_settings
object
- Stack ensemble settings for stack ensemble run.
- allowedTrainingAlgorithms
List<String | "AutoArima" | "Prophet" | "Naive" | "SeasonalNaive" | "Average" | "SeasonalAverage" | "ExponentialSmoothing" | "Arimax" | "TCNForecaster" | "ElasticNet" | "GradientBoosting" | "DecisionTree" | "KNN" | "LassoLars" | "SGD" | "RandomForest" | "ExtremeRandomTrees" | "LightGBM" | "XGBoostRegressor">
- Allowed models for forecasting task.
- blockedTrainingAlgorithms
List<String | "AutoArima" | "Prophet" | "Naive" | "SeasonalNaive" | "Average" | "SeasonalAverage" | "ExponentialSmoothing" | "Arimax" | "TCNForecaster" | "ElasticNet" | "GradientBoosting" | "DecisionTree" | "KNN" | "LassoLars" | "SGD" | "RandomForest" | "ExtremeRandomTrees" | "LightGBM" | "XGBoostRegressor">
- Blocked models for forecasting task.
- enableDnnTraining
Boolean
- Enable recommendation of DNN models.
- enableModelExplainability
Boolean
- Flag to turn on explainability on best model.
- enableOnnxCompatibleModels
Boolean
- Flag for enabling onnx compatible models.
- enableStackEnsemble
Boolean
- Enable stack ensemble run.
- enableVoteEnsemble
Boolean
- Enable voting ensemble run.
- ensembleModelDownloadTimeout
String
- 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
Property Map
- Stack ensemble settings for stack ensemble run.
ForecastingTrainingSettingsResponse
Forecasting Training related configuration.
Goal
[Required] Defines supported metric goals for hyperparameter tuning
| Name | Value |
|---|
| Minimize | Minimize |
| Maximize | Maximize |
| Name | Value |
|---|
| GoalMinimize | Minimize |
| GoalMaximize | Maximize |
| Name | Value |
|---|
| Minimize | Minimize |
| Maximize | Maximize |
| Name | Value |
|---|
| Minimize | Minimize |
| Maximize | Maximize |
| Name | Value |
|---|
| MINIMIZE | Minimize |
| MAXIMIZE | Maximize |
GridSamplingAlgorithm, GridSamplingAlgorithmArgs
Defines a Sampling Algorithm that exhaustively generates every value combination in the space
GridSamplingAlgorithmResponse
Defines a Sampling Algorithm that exhaustively generates every value combination in the space
ImageClassification, ImageClassificationArgs
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, ImageClassificationMultilabelArgs
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'.
ImageClassificationMultilabelResponse
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'.
ImageClassificationResponse
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'.
ImageInstanceSegmentation, ImageInstanceSegmentationArgs
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.
ImageInstanceSegmentationResponse
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.
ImageLimitSettings, ImageLimitSettingsArgs
Limit settings for the AutoML job.
ImageLimitSettingsResponse
Limit settings for the AutoML job.
ImageModelDistributionSettingsClassification, ImageModelDistributionSettingsClassificationArgs
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.
- 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.
- 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.
- ams_gradient
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.
- early_stopping
string
- Enable early stopping logic during training.
- early_stopping_delay
string
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- early_stopping_patience
string
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- enable_onnx_normalization
string
- Enable normalization when exporting ONNX model.
- evaluation_frequency
string
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradient_accumulation_step
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.
- layers_to_freeze
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.
- learning_rate
string
- Initial learning rate. Must be a float in the range [0, 1].
- learning_rate_scheduler
string
- Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'.
- model_name
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'.
- number_of_epochs
string
- Number of training epochs. Must be a positive integer.
- number_of_workers
string
- Number of data loader workers. Must be a non-negative integer.
- optimizer
string
- Type of optimizer. Must be either 'sgd', 'adam', or 'adamw'.
- random_seed
string
- Random seed to be used when using deterministic training.
- step_lr_gamma
string
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- step_lr_step_size
string
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- training_batch_size
string
- Training batch size. Must be a positive integer.
- training_crop_size
string
- Image crop size that is input to the neural network for the training dataset. Must be a positive integer.
- validation_batch_size
string
- Validation batch size. Must be a positive integer.
- validation_crop_size
string
- Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.
- validation_resize_size
string
- Image size to which to resize before cropping for validation dataset. Must be a positive integer.
- warmup_cosine_lr_cycles
string
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmup_cosine_lr_warmup_epochs
string
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- weight_decay
string
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- weighted_loss
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.
- 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.
- 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.
- ams_gradient
str
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- augmentations
str
- Settings for using Augmentations.
- beta1
str
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- beta2
str
- Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- distributed
str
- Whether to use distributer training.
- early_stopping
str
- Enable early stopping logic during training.
- early_stopping_delay
str
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- early_stopping_patience
str
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- enable_onnx_normalization
str
- Enable normalization when exporting ONNX model.
- evaluation_frequency
str
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradient_accumulation_step
str
- 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.
- layers_to_freeze
str
- 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.
- learning_rate
str
- Initial learning rate. Must be a float in the range [0, 1].
- learning_rate_scheduler
str
- Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'.
- model_name
str
- 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
str
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- nesterov
str
- Enable nesterov when optimizer is 'sgd'.
- number_of_epochs
str
- Number of training epochs. Must be a positive integer.
- number_of_workers
str
- Number of data loader workers. Must be a non-negative integer.
- optimizer
str
- Type of optimizer. Must be either 'sgd', 'adam', or 'adamw'.
- random_seed
str
- Random seed to be used when using deterministic training.
- step_lr_gamma
str
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- step_lr_step_size
str
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- training_batch_size
str
- Training batch size. Must be a positive integer.
- training_crop_size
str
- Image crop size that is input to the neural network for the training dataset. Must be a positive integer.
- validation_batch_size
str
- Validation batch size. Must be a positive integer.
- validation_crop_size
str
- Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.
- validation_resize_size
str
- Image size to which to resize before cropping for validation dataset. Must be a positive integer.
- warmup_cosine_lr_cycles
str
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmup_cosine_lr_warmup_epochs
str
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- weight_decay
str
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- weighted_loss
str
- 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.
- 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.
ImageModelDistributionSettingsClassificationResponse
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.
- 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.
- 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.
- ams_gradient
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.
- early_stopping
string
- Enable early stopping logic during training.
- early_stopping_delay
string
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- early_stopping_patience
string
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- enable_onnx_normalization
string
- Enable normalization when exporting ONNX model.
- evaluation_frequency
string
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradient_accumulation_step
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.
- layers_to_freeze
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.
- learning_rate
string
- Initial learning rate. Must be a float in the range [0, 1].
- learning_rate_scheduler
string
- Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'.
- model_name
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'.
- number_of_epochs
string
- Number of training epochs. Must be a positive integer.
- number_of_workers
string
- Number of data loader workers. Must be a non-negative integer.
- optimizer
string
- Type of optimizer. Must be either 'sgd', 'adam', or 'adamw'.
- random_seed
string
- Random seed to be used when using deterministic training.
- step_lr_gamma
string
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- step_lr_step_size
string
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- training_batch_size
string
- Training batch size. Must be a positive integer.
- training_crop_size
string
- Image crop size that is input to the neural network for the training dataset. Must be a positive integer.
- validation_batch_size
string
- Validation batch size. Must be a positive integer.
- validation_crop_size
string
- Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.
- validation_resize_size
string
- Image size to which to resize before cropping for validation dataset. Must be a positive integer.
- warmup_cosine_lr_cycles
string
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmup_cosine_lr_warmup_epochs
string
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- weight_decay
string
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- weighted_loss
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.
- 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.
- 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.
- ams_gradient
str
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- augmentations
str
- Settings for using Augmentations.
- beta1
str
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- beta2
str
- Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- distributed
str
- Whether to use distributer training.
- early_stopping
str
- Enable early stopping logic during training.
- early_stopping_delay
str
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- early_stopping_patience
str
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- enable_onnx_normalization
str
- Enable normalization when exporting ONNX model.
- evaluation_frequency
str
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradient_accumulation_step
str
- 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.
- layers_to_freeze
str
- 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.
- learning_rate
str
- Initial learning rate. Must be a float in the range [0, 1].
- learning_rate_scheduler
str
- Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'.
- model_name
str
- 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
str
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- nesterov
str
- Enable nesterov when optimizer is 'sgd'.
- number_of_epochs
str
- Number of training epochs. Must be a positive integer.
- number_of_workers
str
- Number of data loader workers. Must be a non-negative integer.
- optimizer
str
- Type of optimizer. Must be either 'sgd', 'adam', or 'adamw'.
- random_seed
str
- Random seed to be used when using deterministic training.
- step_lr_gamma
str
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- step_lr_step_size
str
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- training_batch_size
str
- Training batch size. Must be a positive integer.
- training_crop_size
str
- Image crop size that is input to the neural network for the training dataset. Must be a positive integer.
- validation_batch_size
str
- Validation batch size. Must be a positive integer.
- validation_crop_size
str
- Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.
- validation_resize_size
str
- Image size to which to resize before cropping for validation dataset. Must be a positive integer.
- warmup_cosine_lr_cycles
str
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmup_cosine_lr_warmup_epochs
str
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- weight_decay
str
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- weighted_loss
str
- 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.
- 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, ImageModelDistributionSettingsObjectDetectionArgs
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.
- 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].
- 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].
- ams_gradient
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].
- box_detections_per_image
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.
- box_score_threshold
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.
- early_stopping
string
- Enable early stopping logic during training.
- early_stopping_delay
string
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- early_stopping_patience
string
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- enable_onnx_normalization
string
- Enable normalization when exporting ONNX model.
- evaluation_frequency
string
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradient_accumulation_step
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.
- image_size
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.
- layers_to_freeze
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.
- learning_rate
string
- Initial learning rate. Must be a float in the range [0, 1].
- learning_rate_scheduler
string
- Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'.
- max_size
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.
- min_size
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.
- model_name
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.
- model_size
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].
- multi_scale
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'.
- nms_iou_threshold
string
- IOU threshold used during inference in NMS post processing. Must be float in the range [0, 1].
- number_of_epochs
string
- Number of training epochs. Must be a positive integer.
- number_of_workers
string
- Number of data loader workers. Must be a non-negative integer.
- optimizer
string
- Type of optimizer. Must be either 'sgd', 'adam', or 'adamw'.
- random_seed
string
- Random seed to be used when using deterministic training.
- step_lr_gamma
string
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- step_lr_step_size
string
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- tile_grid_size
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.
- tile_overlap_ratio
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.
- tile_predictions_nms_threshold
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
- training_batch_size
string
- Training batch size. Must be a positive integer.
- validation_batch_size
string
- Validation batch size. Must be a positive integer.
- validation_iou_threshold
string
- IOU threshold to use when computing validation metric. Must be float in the range [0, 1].
- validation_metric_type
string
- Metric computation method to use for validation metrics. Must be 'none', 'coco', 'voc', or 'coco_voc'.
- warmup_cosine_lr_cycles
string
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmup_cosine_lr_warmup_epochs
string
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- weight_decay
string
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- 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].
- 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].
- ams_gradient
str
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- augmentations
str
- Settings for using Augmentations.
- beta1
str
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- beta2
str
- Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- box_detections_per_image
str
- Maximum number of detections per image, for all classes. Must be a positive integer.
Note: This settings is not supported for the 'yolov5' algorithm.
- box_score_threshold
str
- During inference, only return proposals with a classification score greater than
BoxScoreThreshold. Must be a float in the range[0, 1].
- distributed
str
- Whether to use distributer training.
- early_stopping
str
- Enable early stopping logic during training.
- early_stopping_delay
str
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- early_stopping_patience
str
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- enable_onnx_normalization
str
- Enable normalization when exporting ONNX model.
- evaluation_frequency
str
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradient_accumulation_step
str
- 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.
- image_size
str
- 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.
- layers_to_freeze
str
- 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.
- learning_rate
str
- Initial learning rate. Must be a float in the range [0, 1].
- learning_rate_scheduler
str
- Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'.
- max_size
str
- 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.
- min_size
str
- 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.
- model_name
str
- 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.
- model_size
str
- 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
str
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- multi_scale
str
- 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
str
- Enable nesterov when optimizer is 'sgd'.
- nms_iou_threshold
str
- IOU threshold used during inference in NMS post processing. Must be float in the range [0, 1].
- number_of_epochs
str
- Number of training epochs. Must be a positive integer.
- number_of_workers
str
- Number of data loader workers. Must be a non-negative integer.
- optimizer
str
- Type of optimizer. Must be either 'sgd', 'adam', or 'adamw'.
- random_seed
str
- Random seed to be used when using deterministic training.
- step_lr_gamma
str
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- step_lr_step_size
str
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- tile_grid_size
str
- 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.
- tile_overlap_ratio
str
- 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.
- tile_predictions_nms_threshold
str
- 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
- training_batch_size
str
- Training batch size. Must be a positive integer.
- validation_batch_size
str
- Validation batch size. Must be a positive integer.
- validation_iou_threshold
str
- IOU threshold to use when computing validation metric. Must be float in the range [0, 1].
- validation_metric_type
str
- Metric computation method to use for validation metrics. Must be 'none', 'coco', 'voc', or 'coco_voc'.
- warmup_cosine_lr_cycles
str
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmup_cosine_lr_warmup_epochs
str
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- weight_decay
str
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- 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].
ImageModelDistributionSettingsObjectDetectionResponse
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.
- 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].
- 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].
- ams_gradient
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].
- box_detections_per_image
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.
- box_score_threshold
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.
- early_stopping
string
- Enable early stopping logic during training.
- early_stopping_delay
string
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- early_stopping_patience
string
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- enable_onnx_normalization
string
- Enable normalization when exporting ONNX model.
- evaluation_frequency
string
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradient_accumulation_step
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.
- image_size
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.
- layers_to_freeze
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.
- learning_rate
string
- Initial learning rate. Must be a float in the range [0, 1].
- learning_rate_scheduler
string
- Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'.
- max_size
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.
- min_size
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.
- model_name
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.
- model_size
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].
- multi_scale
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'.
- nms_iou_threshold
string
- IOU threshold used during inference in NMS post processing. Must be float in the range [0, 1].
- number_of_epochs
string
- Number of training epochs. Must be a positive integer.
- number_of_workers
string
- Number of data loader workers. Must be a non-negative integer.
- optimizer
string
- Type of optimizer. Must be either 'sgd', 'adam', or 'adamw'.
- random_seed
string
- Random seed to be used when using deterministic training.
- step_lr_gamma
string
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- step_lr_step_size
string
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- tile_grid_size
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.
- tile_overlap_ratio
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.
- tile_predictions_nms_threshold
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
- training_batch_size
string
- Training batch size. Must be a positive integer.
- validation_batch_size
string
- Validation batch size. Must be a positive integer.
- validation_iou_threshold
string
- IOU threshold to use when computing validation metric. Must be float in the range [0, 1].
- validation_metric_type
string
- Metric computation method to use for validation metrics. Must be 'none', 'coco', 'voc', or 'coco_voc'.
- warmup_cosine_lr_cycles
string
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmup_cosine_lr_warmup_epochs
string
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- weight_decay
string
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- 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].
- 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].
- ams_gradient
str
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- augmentations
str
- Settings for using Augmentations.
- beta1
str
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- beta2
str
- Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- box_detections_per_image
str
- Maximum number of detections per image, for all classes. Must be a positive integer.
Note: This settings is not supported for the 'yolov5' algorithm.
- box_score_threshold
str
- During inference, only return proposals with a classification score greater than
BoxScoreThreshold. Must be a float in the range[0, 1].
- distributed
str
- Whether to use distributer training.
- early_stopping
str
- Enable early stopping logic during training.
- early_stopping_delay
str
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- early_stopping_patience
str
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- enable_onnx_normalization
str
- Enable normalization when exporting ONNX model.
- evaluation_frequency
str
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradient_accumulation_step
str
- 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.
- image_size
str
- 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.
- layers_to_freeze
str
- 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.
- learning_rate
str
- Initial learning rate. Must be a float in the range [0, 1].
- learning_rate_scheduler
str
- Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'.
- max_size
str
- 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.
- min_size
str
- 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.
- model_name
str
- 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.
- model_size
str
- 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
str
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- multi_scale
str
- 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
str
- Enable nesterov when optimizer is 'sgd'.
- nms_iou_threshold
str
- IOU threshold used during inference in NMS post processing. Must be float in the range [0, 1].
- number_of_epochs
str
- Number of training epochs. Must be a positive integer.
- number_of_workers
str
- Number of data loader workers. Must be a non-negative integer.
- optimizer
str
- Type of optimizer. Must be either 'sgd', 'adam', or 'adamw'.
- random_seed
str
- Random seed to be used when using deterministic training.
- step_lr_gamma
str
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- step_lr_step_size
str
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- tile_grid_size
str
- 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.
- tile_overlap_ratio
str
- 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.
- tile_predictions_nms_threshold
str
- 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
- training_batch_size
str
- Training batch size. Must be a positive integer.
- validation_batch_size
str
- Validation batch size. Must be a positive integer.
- validation_iou_threshold
str
- IOU threshold to use when computing validation metric. Must be float in the range [0, 1].
- validation_metric_type
str
- Metric computation method to use for validation metrics. Must be 'none', 'coco', 'voc', or 'coco_voc'.
- warmup_cosine_lr_cycles
str
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmup_cosine_lr_warmup_epochs
str
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- weight_decay
str
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- 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, ImageModelSettingsClassificationArgs
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.
- AdvancedSettings
string
- Settings for advanced scenarios.
- AmsGradient
bool
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- Augmentations
string
- Settings for using Augmentations.
- Beta1
double
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- Beta2
double
- Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- CheckpointFrequency
int
- Frequency to store model checkpoints. Must be a positive integer.
- CheckpointModel
Pulumi.AzureNative.MachineLearningServices.Inputs.MLFlowModelJobInput
- The pretrained checkpoint model for incremental training.
- CheckpointRunId
string
- The id of a previous run that has a pretrained checkpoint for incremental training.
- Distributed
bool
- Whether to use distributed training.
- EarlyStopping
bool
- Enable early stopping logic during training.
- EarlyStoppingDelay
int
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- EarlyStoppingPatience
int
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- EnableOnnxNormalization
bool
- Enable normalization when exporting ONNX model.
- EvaluationFrequency
int
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- GradientAccumulationStep
int
- 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
int
- 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
double
- Initial learning rate. Must be a float in the range [0, 1].
- LearningRateScheduler
string | Pulumi.AzureNative.MachineLearningServices.LearningRateScheduler
- Learning rate scheduler enum.
- 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
double
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- Nesterov
bool
- Enable nesterov when optimizer is 'sgd'.
- NumberOfEpochs
int
- Number of training epochs. Must be a positive integer.
- NumberOfWorkers
int
- Number of data loader workers. Must be a non-negative integer.
- Optimizer
string | Pulumi.AzureNative.MachineLearningServices.StochasticOptimizer
- Stochastic optimizer for image models.
- RandomSeed
int
- Random seed to be used when using deterministic training.
- StepLRGamma
double
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- StepLRStepSize
int
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- TrainingBatchSize
int
- Training batch size. Must be a positive integer.
- TrainingCropSize
int
- Image crop size that is input to the neural network for the training dataset. Must be a positive integer.
- ValidationBatchSize
int
- Validation batch size. Must be a positive integer.
- ValidationCropSize
int
- Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.
- ValidationResizeSize
int
- Image size to which to resize before cropping for validation dataset. Must be a positive integer.
- WarmupCosineLRCycles
double
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- WarmupCosineLRWarmupEpochs
int
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- WeightDecay
double
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- WeightedLoss
int
- 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.
- AdvancedSettings
string
- Settings for advanced scenarios.
- AmsGradient
bool
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- Augmentations
string
- Settings for using Augmentations.
- Beta1
float64
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- Beta2
float64
- Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- CheckpointFrequency
int
- Frequency to store model checkpoints. Must be a positive integer.
- CheckpointModel
MLFlowModelJobInput
- The pretrained checkpoint model for incremental training.
- CheckpointRunId
string
- The id of a previous run that has a pretrained checkpoint for incremental training.
- Distributed
bool
- Whether to use distributed training.
- EarlyStopping
bool
- Enable early stopping logic during training.
- EarlyStoppingDelay
int
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- EarlyStoppingPatience
int
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- EnableOnnxNormalization
bool
- Enable normalization when exporting ONNX model.
- EvaluationFrequency
int
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- GradientAccumulationStep
int
- 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
int
- 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
float64
- Initial learning rate. Must be a float in the range [0, 1].
- LearningRateScheduler
string | LearningRateScheduler
- Learning rate scheduler enum.
- 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
float64
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- Nesterov
bool
- Enable nesterov when optimizer is 'sgd'.
- NumberOfEpochs
int
- Number of training epochs. Must be a positive integer.
- NumberOfWorkers
int
- Number of data loader workers. Must be a non-negative integer.
- Optimizer
string | StochasticOptimizer
- Stochastic optimizer for image models.
- RandomSeed
int
- Random seed to be used when using deterministic training.
- StepLRGamma
float64
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- StepLRStepSize
int
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- TrainingBatchSize
int
- Training batch size. Must be a positive integer.
- TrainingCropSize
int
- Image crop size that is input to the neural network for the training dataset. Must be a positive integer.
- ValidationBatchSize
int
- Validation batch size. Must be a positive integer.
- ValidationCropSize
int
- Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.
- ValidationResizeSize
int
- Image size to which to resize before cropping for validation dataset. Must be a positive integer.
- WarmupCosineLRCycles
float64
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- WarmupCosineLRWarmupEpochs
int
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- WeightDecay
float64
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- WeightedLoss
int
- 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.
- advanced_settings
string
- Settings for advanced scenarios.
- ams_gradient
bool
- 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].
- checkpoint_frequency
number
- Frequency to store model checkpoints. Must be a positive integer.
- checkpoint_model
object
- The pretrained checkpoint model for incremental training.
- checkpoint_run_id
string
- The id of a previous run that has a pretrained checkpoint for incremental training.
- distributed
bool
- Whether to use distributed training.
- early_stopping
bool
- Enable early stopping logic during training.
- early_stopping_delay
number
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- early_stopping_patience
number
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- enable_onnx_normalization
bool
- Enable normalization when exporting ONNX model.
- evaluation_frequency
number
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradient_accumulation_step
number
- 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.
- layers_to_freeze
number
- 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.
- learning_rate
number
- Initial learning rate. Must be a float in the range [0, 1].
- learning_rate_scheduler
string | "None" | "WarmupCosine" | "Step"
- Learning rate scheduler enum.
- model_name
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
bool
- Enable nesterov when optimizer is 'sgd'.
- number_of_epochs
number
- Number of training epochs. Must be a positive integer.
- number_of_workers
number
- Number of data loader workers. Must be a non-negative integer.
- optimizer
string | "None" | "Sgd" | "Adam" | "Adamw"
- Stochastic optimizer for image models.
- random_seed
number
- Random seed to be used when using deterministic training.
- step_lr_gamma
number
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- step_lr_step_size
number
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- training_batch_size
number
- Training batch size. Must be a positive integer.
- training_crop_size
number
- Image crop size that is input to the neural network for the training dataset. Must be a positive integer.
- validation_batch_size
number
- Validation batch size. Must be a positive integer.
- validation_crop_size
number
- Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.
- validation_resize_size
number
- Image size to which to resize before cropping for validation dataset. Must be a positive integer.
- warmup_cosine_lr_cycles
number
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmup_cosine_lr_warmup_epochs
number
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- weight_decay
number
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- weighted_loss
number
- 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.
- advancedSettings
String
- Settings for advanced scenarios.
- amsGradient
Boolean
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- augmentations
String
- Settings for using Augmentations.
- beta1
Double
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- beta2
Double
- 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
MLFlowModelJobInput
- 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
Double
- Initial learning rate. Must be a float in the range [0, 1].
- learningRateScheduler
String | LearningRateScheduler
- Learning rate scheduler enum.
- 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
Double
- 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
String | StochasticOptimizer
- Stochastic optimizer for image models.
- randomSeed
Integer
- Random seed to be used when using deterministic training.
- stepLRGamma
Double
- 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
Double
- 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
Double
- 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.
- 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
number
- Frequency to store model checkpoints. Must be a positive integer.
- checkpointModel
MLFlowModelJobInput
- 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
number
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- earlyStoppingPatience
number
- 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
number
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradientAccumulationStep
number
- 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
number
- 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
string | LearningRateScheduler
- Learning rate scheduler enum.
- 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
number
- Number of training epochs. Must be a positive integer.
- numberOfWorkers
number
- Number of data loader workers. Must be a non-negative integer.
- optimizer
string | StochasticOptimizer
- Stochastic optimizer for image models.
- randomSeed
number
- 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
number
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- trainingBatchSize
number
- Training batch size. Must be a positive integer.
- trainingCropSize
number
- Image crop size that is input to the neural network for the training dataset. Must be a positive integer.
- validationBatchSize
number
- Validation batch size. Must be a positive integer.
- validationCropSize
number
- Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.
- validationResizeSize
number
- 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
number
- 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
number
- 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.
- advanced_settings
str
- Settings for advanced scenarios.
- ams_gradient
bool
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- augmentations
str
- Settings for using Augmentations.
- beta1
float
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- beta2
float
- Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- checkpoint_frequency
int
- Frequency to store model checkpoints. Must be a positive integer.
- checkpoint_model
MLFlowModelJobInput
- The pretrained checkpoint model for incremental training.
- checkpoint_run_id
str
- The id of a previous run that has a pretrained checkpoint for incremental training.
- distributed
bool
- Whether to use distributed training.
- early_stopping
bool
- Enable early stopping logic during training.
- early_stopping_delay
int
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- early_stopping_patience
int
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- enable_onnx_normalization
bool
- Enable normalization when exporting ONNX model.
- evaluation_frequency
int
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradient_accumulation_step
int
- 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.
- layers_to_freeze
int
- 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.
- learning_rate
float
- Initial learning rate. Must be a float in the range [0, 1].
- learning_rate_scheduler
str | LearningRateScheduler
- Learning rate scheduler enum.
- model_name
str
- 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
float
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- nesterov
bool
- Enable nesterov when optimizer is 'sgd'.
- number_of_epochs
int
- Number of training epochs. Must be a positive integer.
- number_of_workers
int
- Number of data loader workers. Must be a non-negative integer.
- optimizer
str | StochasticOptimizer
- Stochastic optimizer for image models.
- random_seed
int
- Random seed to be used when using deterministic training.
- step_lr_gamma
float
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- step_lr_step_size
int
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- training_batch_size
int
- Training batch size. Must be a positive integer.
- training_crop_size
int
- Image crop size that is input to the neural network for the training dataset. Must be a positive integer.
- validation_batch_size
int
- Validation batch size. Must be a positive integer.
- validation_crop_size
int
- Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.
- validation_resize_size
int
- Image size to which to resize before cropping for validation dataset. Must be a positive integer.
- warmup_cosine_lr_cycles
float
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmup_cosine_lr_warmup_epochs
int
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- weight_decay
float
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- weighted_loss
int
- 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.
- 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
Number
- Frequency to store model checkpoints. Must be a positive integer.
- checkpointModel
Property Map
- 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
Number
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- earlyStoppingPatience
Number
- 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
Number
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradientAccumulationStep
Number
- 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
Number
- 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
String | "None" | "WarmupCosine" | "Step"
- Learning rate scheduler enum.
- 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
Number
- Number of training epochs. Must be a positive integer.
- numberOfWorkers
Number
- Number of data loader workers. Must be a non-negative integer.
- optimizer
String | "None" | "Sgd" | "Adam" | "Adamw"
- Stochastic optimizer for image models.
- randomSeed
Number
- 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
Number
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- trainingBatchSize
Number
- Training batch size. Must be a positive integer.
- trainingCropSize
Number
- Image crop size that is input to the neural network for the training dataset. Must be a positive integer.
- validationBatchSize
Number
- Validation batch size. Must be a positive integer.
- validationCropSize
Number
- Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.
- validationResizeSize
Number
- 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
Number
- 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
Number
- 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.
ImageModelSettingsClassificationResponse
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.
- AdvancedSettings
string
- Settings for advanced scenarios.
- AmsGradient
bool
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- Augmentations
string
- Settings for using Augmentations.
- Beta1
double
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- Beta2
double
- Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- CheckpointFrequency
int
- Frequency to store model checkpoints. Must be a positive integer.
- CheckpointModel
Pulumi.AzureNative.MachineLearningServices.Inputs.MLFlowModelJobInputResponse
- The pretrained checkpoint model for incremental training.
- CheckpointRunId
string
- The id of a previous run that has a pretrained checkpoint for incremental training.
- Distributed
bool
- Whether to use distributed training.
- EarlyStopping
bool
- Enable early stopping logic during training.
- EarlyStoppingDelay
int
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- EarlyStoppingPatience
int
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- EnableOnnxNormalization
bool
- Enable normalization when exporting ONNX model.
- EvaluationFrequency
int
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- GradientAccumulationStep
int
- 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
int
- 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
double
- Initial learning rate. Must be a float in the range [0, 1].
- LearningRateScheduler
string
- Learning rate scheduler enum.
- 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
double
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- Nesterov
bool
- Enable nesterov when optimizer is 'sgd'.
- NumberOfEpochs
int
- Number of training epochs. Must be a positive integer.
- NumberOfWorkers
int
- Number of data loader workers. Must be a non-negative integer.
- Optimizer
string
- Stochastic optimizer for image models.
- RandomSeed
int
- Random seed to be used when using deterministic training.
- StepLRGamma
double
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- StepLRStepSize
int
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- TrainingBatchSize
int
- Training batch size. Must be a positive integer.
- TrainingCropSize
int
- Image crop size that is input to the neural network for the training dataset. Must be a positive integer.
- ValidationBatchSize
int
- Validation batch size. Must be a positive integer.
- ValidationCropSize
int
- Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.
- ValidationResizeSize
int
- Image size to which to resize before cropping for validation dataset. Must be a positive integer.
- WarmupCosineLRCycles
double
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- WarmupCosineLRWarmupEpochs
int
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- WeightDecay
double
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- WeightedLoss
int
- 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.
- AdvancedSettings
string
- Settings for advanced scenarios.
- AmsGradient
bool
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- Augmentations
string
- Settings for using Augmentations.
- Beta1
float64
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- Beta2
float64
- Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- CheckpointFrequency
int
- Frequency to store model checkpoints. Must be a positive integer.
- CheckpointModel
MLFlowModelJobInputResponse
- The pretrained checkpoint model for incremental training.
- CheckpointRunId
string
- The id of a previous run that has a pretrained checkpoint for incremental training.
- Distributed
bool
- Whether to use distributed training.
- EarlyStopping
bool
- Enable early stopping logic during training.
- EarlyStoppingDelay
int
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- EarlyStoppingPatience
int
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- EnableOnnxNormalization
bool
- Enable normalization when exporting ONNX model.
- EvaluationFrequency
int
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- GradientAccumulationStep
int
- 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
int
- 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
float64
- Initial learning rate. Must be a float in the range [0, 1].
- LearningRateScheduler
string
- Learning rate scheduler enum.
- 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
float64
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- Nesterov
bool
- Enable nesterov when optimizer is 'sgd'.
- NumberOfEpochs
int
- Number of training epochs. Must be a positive integer.
- NumberOfWorkers
int
- Number of data loader workers. Must be a non-negative integer.
- Optimizer
string
- Stochastic optimizer for image models.
- RandomSeed
int
- Random seed to be used when using deterministic training.
- StepLRGamma
float64
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- StepLRStepSize
int
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- TrainingBatchSize
int
- Training batch size. Must be a positive integer.
- TrainingCropSize
int
- Image crop size that is input to the neural network for the training dataset. Must be a positive integer.
- ValidationBatchSize
int
- Validation batch size. Must be a positive integer.
- ValidationCropSize
int
- Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.
- ValidationResizeSize
int
- Image size to which to resize before cropping for validation dataset. Must be a positive integer.
- WarmupCosineLRCycles
float64
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- WarmupCosineLRWarmupEpochs
int
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- WeightDecay
float64
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- WeightedLoss
int
- 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.
- advanced_settings
string
- Settings for advanced scenarios.
- ams_gradient
bool
- 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].
- checkpoint_frequency
number
- Frequency to store model checkpoints. Must be a positive integer.
- checkpoint_model
object
- The pretrained checkpoint model for incremental training.
- checkpoint_run_id
string
- The id of a previous run that has a pretrained checkpoint for incremental training.
- distributed
bool
- Whether to use distributed training.
- early_stopping
bool
- Enable early stopping logic during training.
- early_stopping_delay
number
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- early_stopping_patience
number
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- enable_onnx_normalization
bool
- Enable normalization when exporting ONNX model.
- evaluation_frequency
number
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradient_accumulation_step
number
- 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.
- layers_to_freeze
number
- 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.
- learning_rate
number
- Initial learning rate. Must be a float in the range [0, 1].
- learning_rate_scheduler
string
- Learning rate scheduler enum.
- model_name
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
bool
- Enable nesterov when optimizer is 'sgd'.
- number_of_epochs
number
- Number of training epochs. Must be a positive integer.
- number_of_workers
number
- Number of data loader workers. Must be a non-negative integer.
- optimizer
string
- Stochastic optimizer for image models.
- random_seed
number
- Random seed to be used when using deterministic training.
- step_lr_gamma
number
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- step_lr_step_size
number
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- training_batch_size
number
- Training batch size. Must be a positive integer.
- training_crop_size
number
- Image crop size that is input to the neural network for the training dataset. Must be a positive integer.
- validation_batch_size
number
- Validation batch size. Must be a positive integer.
- validation_crop_size
number
- Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.
- validation_resize_size
number
- Image size to which to resize before cropping for validation dataset. Must be a positive integer.
- warmup_cosine_lr_cycles
number
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmup_cosine_lr_warmup_epochs
number
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- weight_decay
number
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- weighted_loss
number
- 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.
- advancedSettings
String
- Settings for advanced scenarios.
- amsGradient
Boolean
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- augmentations
String
- Settings for using Augmentations.
- beta1
Double
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- beta2
Double
- 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
MLFlowModelJobInputResponse
- 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
Double
- Initial learning rate. Must be a float in the range [0, 1].
- learningRateScheduler
String
- Learning rate scheduler enum.
- 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
Double
- 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
String
- Stochastic optimizer for image models.
- randomSeed
Integer
- Random seed to be used when using deterministic training.
- stepLRGamma
Double
- 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
Double
- 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
Double
- 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.
- 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
number
- Frequency to store model checkpoints. Must be a positive integer.
- checkpointModel
MLFlowModelJobInputResponse
- 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
number
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- earlyStoppingPatience
number
- 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
number
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradientAccumulationStep
number
- 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
number
- 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
string
- Learning rate scheduler enum.
- 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
number
- Number of training epochs. Must be a positive integer.
- numberOfWorkers
number
- Number of data loader workers. Must be a non-negative integer.
- optimizer
string
- Stochastic optimizer for image models.
- randomSeed
number
- 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
number
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- trainingBatchSize
number
- Training batch size. Must be a positive integer.
- trainingCropSize
number
- Image crop size that is input to the neural network for the training dataset. Must be a positive integer.
- validationBatchSize
number
- Validation batch size. Must be a positive integer.
- validationCropSize
number
- Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.
- validationResizeSize
number
- 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
number
- 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
number
- 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.
- advanced_settings
str
- Settings for advanced scenarios.
- ams_gradient
bool
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- augmentations
str
- Settings for using Augmentations.
- beta1
float
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- beta2
float
- Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- checkpoint_frequency
int
- Frequency to store model checkpoints. Must be a positive integer.
- checkpoint_model
MLFlowModelJobInputResponse
- The pretrained checkpoint model for incremental training.
- checkpoint_run_id
str
- The id of a previous run that has a pretrained checkpoint for incremental training.
- distributed
bool
- Whether to use distributed training.
- early_stopping
bool
- Enable early stopping logic during training.
- early_stopping_delay
int
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- early_stopping_patience
int
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- enable_onnx_normalization
bool
- Enable normalization when exporting ONNX model.
- evaluation_frequency
int
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradient_accumulation_step
int
- 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.
- layers_to_freeze
int
- 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.
- learning_rate
float
- Initial learning rate. Must be a float in the range [0, 1].
- learning_rate_scheduler
str
- Learning rate scheduler enum.
- model_name
str
- 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
float
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- nesterov
bool
- Enable nesterov when optimizer is 'sgd'.
- number_of_epochs
int
- Number of training epochs. Must be a positive integer.
- number_of_workers
int
- Number of data loader workers. Must be a non-negative integer.
- optimizer
str
- Stochastic optimizer for image models.
- random_seed
int
- Random seed to be used when using deterministic training.
- step_lr_gamma
float
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- step_lr_step_size
int
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- training_batch_size
int
- Training batch size. Must be a positive integer.
- training_crop_size
int
- Image crop size that is input to the neural network for the training dataset. Must be a positive integer.
- validation_batch_size
int
- Validation batch size. Must be a positive integer.
- validation_crop_size
int
- Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.
- validation_resize_size
int
- Image size to which to resize before cropping for validation dataset. Must be a positive integer.
- warmup_cosine_lr_cycles
float
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmup_cosine_lr_warmup_epochs
int
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- weight_decay
float
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- weighted_loss
int
- 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.
- 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
Number
- Frequency to store model checkpoints. Must be a positive integer.
- checkpointModel
Property Map
- 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
Number
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- earlyStoppingPatience
Number
- 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
Number
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradientAccumulationStep
Number
- 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
Number
- 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
String
- Learning rate scheduler enum.
- 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
Number
- Number of training epochs. Must be a positive integer.
- numberOfWorkers
Number
- Number of data loader workers. Must be a non-negative integer.
- optimizer
String
- Stochastic optimizer for image models.
- randomSeed
Number
- 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
Number
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- trainingBatchSize
Number
- Training batch size. Must be a positive integer.
- trainingCropSize
Number
- Image crop size that is input to the neural network for the training dataset. Must be a positive integer.
- validationBatchSize
Number
- Validation batch size. Must be a positive integer.
- validationCropSize
Number
- Image crop size that is input to the neural network for the validation dataset. Must be a positive integer.
- validationResizeSize
Number
- 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
Number
- 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
Number
- 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, ImageModelSettingsObjectDetectionArgs
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.
- AdvancedSettings
string
- Settings for advanced scenarios.
- AmsGradient
bool
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- Augmentations
string
- Settings for using Augmentations.
- Beta1
double
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- Beta2
double
- Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- BoxDetectionsPerImage
int
- 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
double
- During inference, only return proposals with a classification score greater than
BoxScoreThreshold. Must be a float in the range[0, 1].
- CheckpointFrequency
int
- Frequency to store model checkpoints. Must be a positive integer.
- CheckpointModel
Pulumi.AzureNative.MachineLearningServices.Inputs.MLFlowModelJobInput
- The pretrained checkpoint model for incremental training.
- CheckpointRunId
string
- The id of a previous run that has a pretrained checkpoint for incremental training.
- Distributed
bool
- Whether to use distributed training.
- EarlyStopping
bool
- Enable early stopping logic during training.
- EarlyStoppingDelay
int
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- EarlyStoppingPatience
int
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- EnableOnnxNormalization
bool
- Enable normalization when exporting ONNX model.
- EvaluationFrequency
int
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- GradientAccumulationStep
int
- 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
int
- 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
int
- 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
double
- Initial learning rate. Must be a float in the range [0, 1].
- LearningRateScheduler
string | Pulumi.AzureNative.MachineLearningServices.LearningRateScheduler
- Learning rate scheduler enum.
- MaxSize
int
- 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
int
- 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 | Pulumi.AzureNative.MachineLearningServices.ModelSize
- Image model size.
- Momentum
double
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- MultiScale
bool
- 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
bool
- Enable nesterov when optimizer is 'sgd'.
- NmsIouThreshold
double
- IOU threshold used during inference in NMS post processing. Must be a float in the range [0, 1].
- NumberOfEpochs
int
- Number of training epochs. Must be a positive integer.
- NumberOfWorkers
int
- Number of data loader workers. Must be a non-negative integer.
- Optimizer
string | Pulumi.AzureNative.MachineLearningServices.StochasticOptimizer
- Stochastic optimizer for image models.
- RandomSeed
int
- Random seed to be used when using deterministic training.
- StepLRGamma
double
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- StepLRStepSize
int
- 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
double
- 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
double
- 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
int
- Training batch size. Must be a positive integer.
- ValidationBatchSize
int
- Validation batch size. Must be a positive integer.
- ValidationIouThreshold
double
- IOU threshold to use when computing validation metric. Must be float in the range [0, 1].
- ValidationMetricType
string | Pulumi.AzureNative.MachineLearningServices.ValidationMetricType
- Metric computation method to use for validation metrics in image tasks.
- WarmupCosineLRCycles
double
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- WarmupCosineLRWarmupEpochs
int
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- WeightDecay
double
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- AdvancedSettings
string
- Settings for advanced scenarios.
- AmsGradient
bool
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- Augmentations
string
- Settings for using Augmentations.
- Beta1
float64
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- Beta2
float64
- Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- BoxDetectionsPerImage
int
- 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
float64
- During inference, only return proposals with a classification score greater than
BoxScoreThreshold. Must be a float in the range[0, 1].
- CheckpointFrequency
int
- Frequency to store model checkpoints. Must be a positive integer.
- CheckpointModel
MLFlowModelJobInput
- The pretrained checkpoint model for incremental training.
- CheckpointRunId
string
- The id of a previous run that has a pretrained checkpoint for incremental training.
- Distributed
bool
- Whether to use distributed training.
- EarlyStopping
bool
- Enable early stopping logic during training.
- EarlyStoppingDelay
int
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- EarlyStoppingPatience
int
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- EnableOnnxNormalization
bool
- Enable normalization when exporting ONNX model.
- EvaluationFrequency
int
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- GradientAccumulationStep
int
- 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
int
- 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
int
- 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
float64
- Initial learning rate. Must be a float in the range [0, 1].
- LearningRateScheduler
string | LearningRateScheduler
- Learning rate scheduler enum.
- MaxSize
int
- 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
int
- 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 | ModelSize
- Image model size.
- Momentum
float64
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- MultiScale
bool
- 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
bool
- Enable nesterov when optimizer is 'sgd'.
- NmsIouThreshold
float64
- IOU threshold used during inference in NMS post processing. Must be a float in the range [0, 1].
- NumberOfEpochs
int
- Number of training epochs. Must be a positive integer.
- NumberOfWorkers
int
- Number of data loader workers. Must be a non-negative integer.
- Optimizer
string | StochasticOptimizer
- Stochastic optimizer for image models.
- RandomSeed
int
- Random seed to be used when using deterministic training.
- StepLRGamma
float64
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- StepLRStepSize
int
- 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
float64
- 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
float64
- 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
int
- Training batch size. Must be a positive integer.
- ValidationBatchSize
int
- Validation batch size. Must be a positive integer.
- ValidationIouThreshold
float64
- IOU threshold to use when computing validation metric. Must be float in the range [0, 1].
- ValidationMetricType
string | ValidationMetricType
- Metric computation method to use for validation metrics in image tasks.
- WarmupCosineLRCycles
float64
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- WarmupCosineLRWarmupEpochs
int
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- WeightDecay
float64
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- advanced_settings
string
- Settings for advanced scenarios.
- ams_gradient
bool
- 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].
- box_detections_per_image
number
- Maximum number of detections per image, for all classes. Must be a positive integer.
Note: This settings is not supported for the 'yolov5' algorithm.
- box_score_threshold
number
- During inference, only return proposals with a classification score greater than
BoxScoreThreshold. Must be a float in the range[0, 1].
- checkpoint_frequency
number
- Frequency to store model checkpoints. Must be a positive integer.
- checkpoint_model
object
- The pretrained checkpoint model for incremental training.
- checkpoint_run_id
string
- The id of a previous run that has a pretrained checkpoint for incremental training.
- distributed
bool
- Whether to use distributed training.
- early_stopping
bool
- Enable early stopping logic during training.
- early_stopping_delay
number
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- early_stopping_patience
number
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- enable_onnx_normalization
bool
- Enable normalization when exporting ONNX model.
- evaluation_frequency
number
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradient_accumulation_step
number
- 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.
- image_size
number
- 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.
- layers_to_freeze
number
- 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.
- learning_rate
number
- Initial learning rate. Must be a float in the range [0, 1].
- learning_rate_scheduler
string | "None" | "WarmupCosine" | "Step"
- Learning rate scheduler enum.
- max_size
number
- 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.
- min_size
number
- 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.
- model_name
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.
- model_size
string | "None" | "Small" | "Medium" | "Large" | "ExtraLarge"
- Image model size.
- momentum
number
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- multi_scale
bool
- 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
bool
- Enable nesterov when optimizer is 'sgd'.
- nms_iou_threshold
number
- IOU threshold used during inference in NMS post processing. Must be a float in the range [0, 1].
- number_of_epochs
number
- Number of training epochs. Must be a positive integer.
- number_of_workers
number
- Number of data loader workers. Must be a non-negative integer.
- optimizer
string | "None" | "Sgd" | "Adam" | "Adamw"
- Stochastic optimizer for image models.
- random_seed
number
- Random seed to be used when using deterministic training.
- step_lr_gamma
number
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- step_lr_step_size
number
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- tile_grid_size
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.
- tile_overlap_ratio
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.
- tile_predictions_nms_threshold
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.
- training_batch_size
number
- Training batch size. Must be a positive integer.
- validation_batch_size
number
- Validation batch size. Must be a positive integer.
- validation_iou_threshold
number
- IOU threshold to use when computing validation metric. Must be float in the range [0, 1].
- validation_metric_type
string | "None" | "Coco" | "Voc" | "CocoVoc"
- Metric computation method to use for validation metrics in image tasks.
- warmup_cosine_lr_cycles
number
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmup_cosine_lr_warmup_epochs
number
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- weight_decay
number
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- advancedSettings
String
- Settings for advanced scenarios.
- amsGradient
Boolean
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- augmentations
String
- Settings for using Augmentations.
- beta1
Double
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- beta2
Double
- 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
Double
- 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
MLFlowModelJobInput
- 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
Double
- Initial learning rate. Must be a float in the range [0, 1].
- learningRateScheduler
String | LearningRateScheduler
- Learning rate scheduler enum.
- 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
String | ModelSize
- Image model size.
- momentum
Double
- 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
Double
- 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
String | StochasticOptimizer
- Stochastic optimizer for image models.
- randomSeed
Integer
- Random seed to be used when using deterministic training.
- stepLRGamma
Double
- 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
Double
- 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
Double
- 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
Double
- IOU threshold to use when computing validation metric. Must be float in the range [0, 1].
- validationMetricType
String | ValidationMetricType
- Metric computation method to use for validation metrics in image tasks.
- warmupCosineLRCycles
Double
- 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
Double
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- 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
number
- 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
number
- Frequency to store model checkpoints. Must be a positive integer.
- checkpointModel
MLFlowModelJobInput
- 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
number
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- earlyStoppingPatience
number
- 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
number
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradientAccumulationStep
number
- 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
number
- 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
number
- 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
string | LearningRateScheduler
- Learning rate scheduler enum.
- maxSize
number
- 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
number
- 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 | ModelSize
- Image model size.
- 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
number
- Number of training epochs. Must be a positive integer.
- numberOfWorkers
number
- Number of data loader workers. Must be a non-negative integer.
- optimizer
string | StochasticOptimizer
- Stochastic optimizer for image models.
- randomSeed
number
- 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
number
- 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
number
- Training batch size. Must be a positive integer.
- validationBatchSize
number
- 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
string | ValidationMetricType
- Metric computation method to use for validation metrics in image tasks.
- warmupCosineLRCycles
number
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmupCosineLRWarmupEpochs
number
- 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].
- advanced_settings
str
- Settings for advanced scenarios.
- ams_gradient
bool
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- augmentations
str
- Settings for using Augmentations.
- beta1
float
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- beta2
float
- Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- box_detections_per_image
int
- Maximum number of detections per image, for all classes. Must be a positive integer.
Note: This settings is not supported for the 'yolov5' algorithm.
- box_score_threshold
float
- During inference, only return proposals with a classification score greater than
BoxScoreThreshold. Must be a float in the range[0, 1].
- checkpoint_frequency
int
- Frequency to store model checkpoints. Must be a positive integer.
- checkpoint_model
MLFlowModelJobInput
- The pretrained checkpoint model for incremental training.
- checkpoint_run_id
str
- The id of a previous run that has a pretrained checkpoint for incremental training.
- distributed
bool
- Whether to use distributed training.
- early_stopping
bool
- Enable early stopping logic during training.
- early_stopping_delay
int
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- early_stopping_patience
int
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- enable_onnx_normalization
bool
- Enable normalization when exporting ONNX model.
- evaluation_frequency
int
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradient_accumulation_step
int
- 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.
- image_size
int
- 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.
- layers_to_freeze
int
- 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.
- learning_rate
float
- Initial learning rate. Must be a float in the range [0, 1].
- learning_rate_scheduler
str | LearningRateScheduler
- Learning rate scheduler enum.
- max_size
int
- 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.
- min_size
int
- 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.
- model_name
str
- 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.
- model_size
str | ModelSize
- Image model size.
- momentum
float
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- multi_scale
bool
- 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
bool
- Enable nesterov when optimizer is 'sgd'.
- nms_iou_threshold
float
- IOU threshold used during inference in NMS post processing. Must be a float in the range [0, 1].
- number_of_epochs
int
- Number of training epochs. Must be a positive integer.
- number_of_workers
int
- Number of data loader workers. Must be a non-negative integer.
- optimizer
str | StochasticOptimizer
- Stochastic optimizer for image models.
- random_seed
int
- Random seed to be used when using deterministic training.
- step_lr_gamma
float
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- step_lr_step_size
int
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- tile_grid_size
str
- 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.
- tile_overlap_ratio
float
- 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.
- tile_predictions_nms_threshold
float
- 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.
- training_batch_size
int
- Training batch size. Must be a positive integer.
- validation_batch_size
int
- Validation batch size. Must be a positive integer.
- validation_iou_threshold
float
- IOU threshold to use when computing validation metric. Must be float in the range [0, 1].
- validation_metric_type
str | ValidationMetricType
- Metric computation method to use for validation metrics in image tasks.
- warmup_cosine_lr_cycles
float
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmup_cosine_lr_warmup_epochs
int
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- weight_decay
float
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- 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
Number
- 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
Number
- Frequency to store model checkpoints. Must be a positive integer.
- checkpointModel
Property Map
- 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
Number
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- earlyStoppingPatience
Number
- 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
Number
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradientAccumulationStep
Number
- 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
Number
- 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
Number
- 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
String | "None" | "WarmupCosine" | "Step"
- Learning rate scheduler enum.
- maxSize
Number
- 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
Number
- 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 | "None" | "Small" | "Medium" | "Large" | "ExtraLarge"
- Image model size.
- 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
Number
- Number of training epochs. Must be a positive integer.
- numberOfWorkers
Number
- Number of data loader workers. Must be a non-negative integer.
- optimizer
String | "None" | "Sgd" | "Adam" | "Adamw"
- Stochastic optimizer for image models.
- randomSeed
Number
- 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
Number
- 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
Number
- Training batch size. Must be a positive integer.
- validationBatchSize
Number
- 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
String | "None" | "Coco" | "Voc" | "CocoVoc"
- Metric computation method to use for validation metrics in image tasks.
- warmupCosineLRCycles
Number
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmupCosineLRWarmupEpochs
Number
- 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].
ImageModelSettingsObjectDetectionResponse
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.
- AdvancedSettings
string
- Settings for advanced scenarios.
- AmsGradient
bool
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- Augmentations
string
- Settings for using Augmentations.
- Beta1
double
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- Beta2
double
- Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- BoxDetectionsPerImage
int
- 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
double
- During inference, only return proposals with a classification score greater than
BoxScoreThreshold. Must be a float in the range[0, 1].
- CheckpointFrequency
int
- Frequency to store model checkpoints. Must be a positive integer.
- CheckpointModel
Pulumi.AzureNative.MachineLearningServices.Inputs.MLFlowModelJobInputResponse
- The pretrained checkpoint model for incremental training.
- CheckpointRunId
string
- The id of a previous run that has a pretrained checkpoint for incremental training.
- Distributed
bool
- Whether to use distributed training.
- EarlyStopping
bool
- Enable early stopping logic during training.
- EarlyStoppingDelay
int
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- EarlyStoppingPatience
int
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- EnableOnnxNormalization
bool
- Enable normalization when exporting ONNX model.
- EvaluationFrequency
int
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- GradientAccumulationStep
int
- 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
int
- 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
int
- 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
double
- Initial learning rate. Must be a float in the range [0, 1].
- LearningRateScheduler
string
- Learning rate scheduler enum.
- MaxSize
int
- 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
int
- 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
- Image model size.
- Momentum
double
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- MultiScale
bool
- 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
bool
- Enable nesterov when optimizer is 'sgd'.
- NmsIouThreshold
double
- IOU threshold used during inference in NMS post processing. Must be a float in the range [0, 1].
- NumberOfEpochs
int
- Number of training epochs. Must be a positive integer.
- NumberOfWorkers
int
- Number of data loader workers. Must be a non-negative integer.
- Optimizer
string
- Stochastic optimizer for image models.
- RandomSeed
int
- Random seed to be used when using deterministic training.
- StepLRGamma
double
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- StepLRStepSize
int
- 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
double
- 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
double
- 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
int
- Training batch size. Must be a positive integer.
- ValidationBatchSize
int
- Validation batch size. Must be a positive integer.
- ValidationIouThreshold
double
- 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 in image tasks.
- WarmupCosineLRCycles
double
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- WarmupCosineLRWarmupEpochs
int
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- WeightDecay
double
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- AdvancedSettings
string
- Settings for advanced scenarios.
- AmsGradient
bool
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- Augmentations
string
- Settings for using Augmentations.
- Beta1
float64
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- Beta2
float64
- Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- BoxDetectionsPerImage
int
- 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
float64
- During inference, only return proposals with a classification score greater than
BoxScoreThreshold. Must be a float in the range[0, 1].
- CheckpointFrequency
int
- Frequency to store model checkpoints. Must be a positive integer.
- CheckpointModel
MLFlowModelJobInputResponse
- The pretrained checkpoint model for incremental training.
- CheckpointRunId
string
- The id of a previous run that has a pretrained checkpoint for incremental training.
- Distributed
bool
- Whether to use distributed training.
- EarlyStopping
bool
- Enable early stopping logic during training.
- EarlyStoppingDelay
int
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- EarlyStoppingPatience
int
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- EnableOnnxNormalization
bool
- Enable normalization when exporting ONNX model.
- EvaluationFrequency
int
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- GradientAccumulationStep
int
- 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
int
- 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
int
- 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
float64
- Initial learning rate. Must be a float in the range [0, 1].
- LearningRateScheduler
string
- Learning rate scheduler enum.
- MaxSize
int
- 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
int
- 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
- Image model size.
- Momentum
float64
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- MultiScale
bool
- 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
bool
- Enable nesterov when optimizer is 'sgd'.
- NmsIouThreshold
float64
- IOU threshold used during inference in NMS post processing. Must be a float in the range [0, 1].
- NumberOfEpochs
int
- Number of training epochs. Must be a positive integer.
- NumberOfWorkers
int
- Number of data loader workers. Must be a non-negative integer.
- Optimizer
string
- Stochastic optimizer for image models.
- RandomSeed
int
- Random seed to be used when using deterministic training.
- StepLRGamma
float64
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- StepLRStepSize
int
- 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
float64
- 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
float64
- 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
int
- Training batch size. Must be a positive integer.
- ValidationBatchSize
int
- Validation batch size. Must be a positive integer.
- ValidationIouThreshold
float64
- 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 in image tasks.
- WarmupCosineLRCycles
float64
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- WarmupCosineLRWarmupEpochs
int
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- WeightDecay
float64
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- advanced_settings
string
- Settings for advanced scenarios.
- ams_gradient
bool
- 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].
- box_detections_per_image
number
- Maximum number of detections per image, for all classes. Must be a positive integer.
Note: This settings is not supported for the 'yolov5' algorithm.
- box_score_threshold
number
- During inference, only return proposals with a classification score greater than
BoxScoreThreshold. Must be a float in the range[0, 1].
- checkpoint_frequency
number
- Frequency to store model checkpoints. Must be a positive integer.
- checkpoint_model
object
- The pretrained checkpoint model for incremental training.
- checkpoint_run_id
string
- The id of a previous run that has a pretrained checkpoint for incremental training.
- distributed
bool
- Whether to use distributed training.
- early_stopping
bool
- Enable early stopping logic during training.
- early_stopping_delay
number
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- early_stopping_patience
number
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- enable_onnx_normalization
bool
- Enable normalization when exporting ONNX model.
- evaluation_frequency
number
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradient_accumulation_step
number
- 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.
- image_size
number
- 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.
- layers_to_freeze
number
- 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.
- learning_rate
number
- Initial learning rate. Must be a float in the range [0, 1].
- learning_rate_scheduler
string
- Learning rate scheduler enum.
- max_size
number
- 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.
- min_size
number
- 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.
- model_name
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.
- model_size
string
- Image model size.
- momentum
number
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- multi_scale
bool
- 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
bool
- Enable nesterov when optimizer is 'sgd'.
- nms_iou_threshold
number
- IOU threshold used during inference in NMS post processing. Must be a float in the range [0, 1].
- number_of_epochs
number
- Number of training epochs. Must be a positive integer.
- number_of_workers
number
- Number of data loader workers. Must be a non-negative integer.
- optimizer
string
- Stochastic optimizer for image models.
- random_seed
number
- Random seed to be used when using deterministic training.
- step_lr_gamma
number
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- step_lr_step_size
number
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- tile_grid_size
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.
- tile_overlap_ratio
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.
- tile_predictions_nms_threshold
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.
- training_batch_size
number
- Training batch size. Must be a positive integer.
- validation_batch_size
number
- Validation batch size. Must be a positive integer.
- validation_iou_threshold
number
- IOU threshold to use when computing validation metric. Must be float in the range [0, 1].
- validation_metric_type
string
- Metric computation method to use for validation metrics in image tasks.
- warmup_cosine_lr_cycles
number
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmup_cosine_lr_warmup_epochs
number
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- weight_decay
number
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- advancedSettings
String
- Settings for advanced scenarios.
- amsGradient
Boolean
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- augmentations
String
- Settings for using Augmentations.
- beta1
Double
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- beta2
Double
- 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
Double
- 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
MLFlowModelJobInputResponse
- 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
Double
- Initial learning rate. Must be a float in the range [0, 1].
- learningRateScheduler
String
- Learning rate scheduler enum.
- 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
String
- Image model size.
- momentum
Double
- 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
Double
- 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
String
- Stochastic optimizer for image models.
- randomSeed
Integer
- Random seed to be used when using deterministic training.
- stepLRGamma
Double
- 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
Double
- 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
Double
- 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
Double
- 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 in image tasks.
- warmupCosineLRCycles
Double
- 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
Double
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- 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
number
- 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
number
- Frequency to store model checkpoints. Must be a positive integer.
- checkpointModel
MLFlowModelJobInputResponse
- 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
number
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- earlyStoppingPatience
number
- 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
number
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradientAccumulationStep
number
- 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
number
- 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
number
- 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
string
- Learning rate scheduler enum.
- maxSize
number
- 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
number
- 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
- Image model size.
- 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
number
- Number of training epochs. Must be a positive integer.
- numberOfWorkers
number
- Number of data loader workers. Must be a non-negative integer.
- optimizer
string
- Stochastic optimizer for image models.
- randomSeed
number
- 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
number
- 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
number
- Training batch size. Must be a positive integer.
- validationBatchSize
number
- 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
string
- Metric computation method to use for validation metrics in image tasks.
- warmupCosineLRCycles
number
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmupCosineLRWarmupEpochs
number
- 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].
- advanced_settings
str
- Settings for advanced scenarios.
- ams_gradient
bool
- Enable AMSGrad when optimizer is 'adam' or 'adamw'.
- augmentations
str
- Settings for using Augmentations.
- beta1
float
- Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- beta2
float
- Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
- box_detections_per_image
int
- Maximum number of detections per image, for all classes. Must be a positive integer.
Note: This settings is not supported for the 'yolov5' algorithm.
- box_score_threshold
float
- During inference, only return proposals with a classification score greater than
BoxScoreThreshold. Must be a float in the range[0, 1].
- checkpoint_frequency
int
- Frequency to store model checkpoints. Must be a positive integer.
- checkpoint_model
MLFlowModelJobInputResponse
- The pretrained checkpoint model for incremental training.
- checkpoint_run_id
str
- The id of a previous run that has a pretrained checkpoint for incremental training.
- distributed
bool
- Whether to use distributed training.
- early_stopping
bool
- Enable early stopping logic during training.
- early_stopping_delay
int
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- early_stopping_patience
int
- Minimum number of epochs or validation evaluations with no primary metric improvement before
the run is stopped. Must be a positive integer.
- enable_onnx_normalization
bool
- Enable normalization when exporting ONNX model.
- evaluation_frequency
int
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradient_accumulation_step
int
- 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.
- image_size
int
- 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.
- layers_to_freeze
int
- 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.
- learning_rate
float
- Initial learning rate. Must be a float in the range [0, 1].
- learning_rate_scheduler
str
- Learning rate scheduler enum.
- max_size
int
- 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.
- min_size
int
- 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.
- model_name
str
- 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.
- model_size
str
- Image model size.
- momentum
float
- Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
- multi_scale
bool
- 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
bool
- Enable nesterov when optimizer is 'sgd'.
- nms_iou_threshold
float
- IOU threshold used during inference in NMS post processing. Must be a float in the range [0, 1].
- number_of_epochs
int
- Number of training epochs. Must be a positive integer.
- number_of_workers
int
- Number of data loader workers. Must be a non-negative integer.
- optimizer
str
- Stochastic optimizer for image models.
- random_seed
int
- Random seed to be used when using deterministic training.
- step_lr_gamma
float
- Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
- step_lr_step_size
int
- Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
- tile_grid_size
str
- 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.
- tile_overlap_ratio
float
- 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.
- tile_predictions_nms_threshold
float
- 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.
- training_batch_size
int
- Training batch size. Must be a positive integer.
- validation_batch_size
int
- Validation batch size. Must be a positive integer.
- validation_iou_threshold
float
- IOU threshold to use when computing validation metric. Must be float in the range [0, 1].
- validation_metric_type
str
- Metric computation method to use for validation metrics in image tasks.
- warmup_cosine_lr_cycles
float
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmup_cosine_lr_warmup_epochs
int
- Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
- weight_decay
float
- Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
- 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
Number
- 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
Number
- Frequency to store model checkpoints. Must be a positive integer.
- checkpointModel
Property Map
- 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
Number
- Minimum number of epochs or validation evaluations to wait before primary metric improvement
is tracked for early stopping. Must be a positive integer.
- earlyStoppingPatience
Number
- 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
Number
- Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
- gradientAccumulationStep
Number
- 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
Number
- 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
Number
- 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
String
- Learning rate scheduler enum.
- maxSize
Number
- 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
Number
- 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
- Image model size.
- 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
Number
- Number of training epochs. Must be a positive integer.
- numberOfWorkers
Number
- Number of data loader workers. Must be a non-negative integer.
- optimizer
String
- Stochastic optimizer for image models.
- randomSeed
Number
- 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
Number
- 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
Number
- Training batch size. Must be a positive integer.
- validationBatchSize
Number
- 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
String
- Metric computation method to use for validation metrics in image tasks.
- warmupCosineLRCycles
Number
- Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
- warmupCosineLRWarmupEpochs
Number
- 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, ImageObjectDetectionArgs
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.
ImageObjectDetectionResponse
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.
ImageSweepSettings, ImageSweepSettingsArgs
Model sweeping and hyperparameter sweeping related settings.
ImageSweepSettingsResponse
Model sweeping and hyperparameter sweeping related settings.
Enum to determine the input data delivery mode.
| Name | Value |
|---|
| ReadOnlyMount | ReadOnlyMount |
| ReadWriteMount | ReadWriteMount |
| Download | Download |
| Direct | Direct |
| EvalMount | EvalMount |
| EvalDownload | EvalDownload |
| Name | Value |
|---|
| InputDeliveryModeReadOnlyMount | ReadOnlyMount |
| InputDeliveryModeReadWriteMount | ReadWriteMount |
| InputDeliveryModeDownload | Download |
| InputDeliveryModeDirect | Direct |
| InputDeliveryModeEvalMount | EvalMount |
| InputDeliveryModeEvalDownload | EvalDownload |
| Value |
|---|
ReadOnlyMount |
ReadWriteMount |
Download |
Direct |
EvalMount |
EvalDownload |
| Name | Value |
|---|
| ReadOnlyMount | ReadOnlyMount |
| ReadWriteMount | ReadWriteMount |
| Download | Download |
| Direct | Direct |
| EvalMount | EvalMount |
| EvalDownload | EvalDownload |
| Name | Value |
|---|
| ReadOnlyMount | ReadOnlyMount |
| ReadWriteMount | ReadWriteMount |
| Download | Download |
| Direct | Direct |
| EvalMount | EvalMount |
| EvalDownload | EvalDownload |
| Name | Value |
|---|
| READ_ONLY_MOUNT | ReadOnlyMount |
| READ_WRITE_MOUNT | ReadWriteMount |
| DOWNLOAD | Download |
| DIRECT | Direct |
| EVAL_MOUNT | EvalMount |
| EVAL_DOWNLOAD | EvalDownload |
| Value |
|---|
ReadOnlyMount |
ReadWriteMount |
Download |
Direct |
EvalMount |
EvalDownload |
InstanceSegmentationPrimaryMetrics
Primary metrics for InstanceSegmentation tasks.
| Name | Value | Description |
|---|
| MeanAveragePrecision | MeanAveragePrecision | Mean Average Precision (MAP) is the average of AP (Average Precision).
AP is calculated for each class and averaged to get the MAP. |
| Name | Value | Description |
|---|
| InstanceSegmentationPrimaryMetricsMeanAveragePrecision | MeanAveragePrecision | Mean Average Precision (MAP) is the average of AP (Average Precision).
AP is calculated for each class and averaged to get the MAP. |
| Value | Description |
|---|
MeanAveragePrecision | Mean Average Precision (MAP) is the average of AP (Average Precision).
AP is calculated for each class and averaged to get the MAP. |
| Name | Value | Description |
|---|
| MeanAveragePrecision | MeanAveragePrecision | Mean Average Precision (MAP) is the average of AP (Average Precision).
AP is calculated for each class and averaged to get the MAP. |
| Name | Value | Description |
|---|
| MeanAveragePrecision | MeanAveragePrecision | Mean Average Precision (MAP) is the average of AP (Average Precision).
AP is calculated for each class and averaged to get the MAP. |
| Name | Value | Description |
|---|
| MEAN_AVERAGE_PRECISION | MeanAveragePrecision | Mean Average Precision (MAP) is the average of AP (Average Precision).
AP is calculated for each class and averaged to get the MAP. |
| Value | Description |
|---|
MeanAveragePrecision | Mean Average Precision (MAP) is the average of AP (Average Precision).
AP is calculated for each class and averaged to get the MAP. |
[Required] Specifies the type of job.
| Name | Value |
|---|
| Literal | literal |
| Uri_file | uri_file |
| Uri_folder | uri_folder |
| Mltable | mltable |
| Custom_model | custom_model |
| Mlflow_model | mlflow_model |
| Triton_model | triton_model |
| Name | Value |
|---|
| JobInputTypeLiteral | literal |
| JobInputType_Uri_file | uri_file |
| JobInputType_Uri_folder | uri_folder |
| JobInputTypeMltable | mltable |
| JobInputType_Custom_model | custom_model |
| JobInputType_Mlflow_model | mlflow_model |
| JobInputType_Triton_model | triton_model |
| Value |
|---|
literal |
uri_file |
uri_folder |
mltable |
custom_model |
mlflow_model |
triton_model |
| Name | Value |
|---|
| Literal | literal |
| Uri_file | uri_file |
| Uri_folder | uri_folder |
| Mltable | mltable |
| Custom_model | custom_model |
| Mlflow_model | mlflow_model |
| Triton_model | triton_model |
| Name | Value |
|---|
| Literal | literal |
| Uri_file | uri_file |
| Uri_folder | uri_folder |
| Mltable | mltable |
| Custom_model | custom_model |
| Mlflow_model | mlflow_model |
| Triton_model | triton_model |
| Name | Value |
|---|
| LITERAL | literal |
| URI_FILE | uri_file |
| URI_FOLDER | uri_folder |
| MLTABLE | mltable |
| CUSTOM_MODEL | custom_model |
| MLFLOW_MODEL | mlflow_model |
| TRITON_MODEL | triton_model |
| Value |
|---|
literal |
uri_file |
uri_folder |
mltable |
custom_model |
mlflow_model |
triton_model |
JobResourceConfiguration, JobResourceConfigurationArgs
- 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.
- DockerArgsList
List<string>
- Extra arguments to pass to the Docker run command, as a collection. 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
int
- 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
Dictionary<string, object>
- Additional properties bag.
- ShmSize
string
- 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).
- 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.
- DockerArgsList
[]string
- Extra arguments to pass to the Docker run command, as a collection. 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
int
- 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
map[string]interface{}
- Additional properties bag.
- ShmSize
string
- 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).
- docker_args
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.
- docker_args_list
list(string)
- Extra arguments to pass to the Docker run command, as a collection. 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.
- instance_count
number
- Optional number of instances or nodes used by the compute target.
- instance_type
string
- Optional type of VM used as supported by the compute target.
- properties
map(any)
- Additional properties bag.
- shm_size
string
- 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).
- 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.
- dockerArgsList
List<String>
- Extra arguments to pass to the Docker run command, as a collection. 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
- 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
Map<String,Object>
- Additional properties bag.
- shmSize
String
- 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).
- 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.
- dockerArgsList
string[]
- Extra arguments to pass to the Docker run command, as a collection. 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
number
- 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
{[key: string]: any}
- Additional properties bag.
- shmSize
string
- 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).
- docker_args
str
- 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.
- docker_args_list
Sequence[str]
- Extra arguments to pass to the Docker run command, as a collection. 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.
- instance_count
int
- Optional number of instances or nodes used by the compute target.
- instance_type
str
- Optional type of VM used as supported by the compute target.
- properties
Mapping[str, Any]
- Additional properties bag.
- shm_size
str
- 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).
- 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.
- dockerArgsList
List<String>
- Extra arguments to pass to the Docker run command, as a collection. 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
Number
- 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
Map<Any>
- Additional properties bag.
- shmSize
String
- 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).
JobResourceConfigurationResponse
- 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.
- DockerArgsList
List<string>
- Extra arguments to pass to the Docker run command, as a collection. 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
int
- 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
Dictionary<string, object>
- Additional properties bag.
- ShmSize
string
- 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).
- 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.
- DockerArgsList
[]string
- Extra arguments to pass to the Docker run command, as a collection. 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
int
- 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
map[string]interface{}
- Additional properties bag.
- ShmSize
string
- 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).
- docker_args
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.
- docker_args_list
list(string)
- Extra arguments to pass to the Docker run command, as a collection. 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.
- instance_count
number
- Optional number of instances or nodes used by the compute target.
- instance_type
string
- Optional type of VM used as supported by the compute target.
- properties
map(any)
- Additional properties bag.
- shm_size
string
- 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).
- 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.
- dockerArgsList
List<String>
- Extra arguments to pass to the Docker run command, as a collection. 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
- 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
Map<String,Object>
- Additional properties bag.
- shmSize
String
- 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).
- 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.
- dockerArgsList
string[]
- Extra arguments to pass to the Docker run command, as a collection. 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
number
- 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
{[key: string]: any}
- Additional properties bag.
- shmSize
string
- 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).
- docker_args
str
- 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.
- docker_args_list
Sequence[str]
- Extra arguments to pass to the Docker run command, as a collection. 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.
- instance_count
int
- Optional number of instances or nodes used by the compute target.
- instance_type
str
- Optional type of VM used as supported by the compute target.
- properties
Mapping[str, Any]
- Additional properties bag.
- shm_size
str
- 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).
- 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.
- dockerArgsList
List<String>
- Extra arguments to pass to the Docker run command, as a collection. 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
Number
- 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
Map<Any>
- Additional properties bag.
- shmSize
String
- 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, JobScheduleActionArgs
JobScheduleActionResponse
JobService, JobServiceArgs
Job endpoint definition
- Endpoint
string
- Url for endpoint.
- JobServiceType
string
- Endpoint type.
- Nodes
AllNodes
- 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
int
- Port for endpoint.
- Properties
map[string]string
- Additional properties to set on the endpoint.
- endpoint
string
- Url for endpoint.
- job_service_type
string
- Endpoint type.
- nodes
object
- 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
number
- Port for endpoint.
- properties
map(string)
- Additional properties to set on the endpoint.
- endpoint
String
- Url for endpoint.
- jobServiceType
String
- Endpoint type.
- nodes
AllNodes
- 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
Map<String,String>
- Additional properties to set on the endpoint.
- endpoint
string
- Url for endpoint.
- jobServiceType
string
- Endpoint type.
- nodes
AllNodes
- 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
number
- Port for endpoint.
- properties
{[key: string]: string}
- Additional properties to set on the endpoint.
- endpoint
str
- Url for endpoint.
- job_service_type
str
- Endpoint type.
- nodes
AllNodes
- 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
int
- Port for endpoint.
- properties
Mapping[str, str]
- Additional properties to set on the endpoint.
- endpoint
String
- Url for endpoint.
- jobServiceType
String
- Endpoint type.
- nodes
Property Map
- 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
Number
- Port for endpoint.
- properties
Map<String>
- Additional properties to set on the endpoint.
JobServiceResponse
Job endpoint definition
- ErrorMessage
string
- Any error in the service.
- Status
string
- Status of endpoint.
- Endpoint
string
- Url for endpoint.
- JobServiceType
string
- Endpoint type.
- Nodes
AllNodesResponse
- 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
int
- Port for endpoint.
- Properties
map[string]string
- Additional properties to set on the endpoint.
- error_message
string
- Any error in the service.
- status
string
- Status of endpoint.
- endpoint
string
- Url for endpoint.
- job_service_type
string
- Endpoint type.
- nodes
object
- 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
number
- Port for endpoint.
- properties
map(string)
- Additional properties to set on the endpoint.
- errorMessage
String
- Any error in the service.
- status
String
- Status of endpoint.
- endpoint
String
- Url for endpoint.
- jobServiceType
String
- Endpoint type.
- nodes
AllNodesResponse
- 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
Map<String,String>
- Additional properties to set on the endpoint.
- errorMessage
string
- Any error in the service.
- status
string
- Status of endpoint.
- endpoint
string
- Url for endpoint.
- jobServiceType
string
- Endpoint type.
- nodes
AllNodesResponse
- 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
number
- Port for endpoint.
- properties
{[key: string]: string}
- Additional properties to set on the endpoint.
- error_message
str
- Any error in the service.
- status
str
- Status of endpoint.
- endpoint
str
- Url for endpoint.
- job_service_type
str
- Endpoint type.
- nodes
AllNodesResponse
- 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
int
- Port for endpoint.
- properties
Mapping[str, str]
- Additional properties to set on the endpoint.
- errorMessage
String
- Any error in the service.
- status
String
- Status of endpoint.
- endpoint
String
- Url for endpoint.
- jobServiceType
String
- Endpoint type.
- nodes
Property Map
- 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
Number
- Port for endpoint.
- properties
Map<String>
- Additional properties to set on the endpoint.
JobTier
Enum to determine the job tier.
| Name | Value |
|---|
| Null | Null |
| Spot | Spot |
| Basic | Basic |
| Standard | Standard |
| Premium | Premium |
| Name | Value |
|---|
| JobTierNull | Null |
| JobTierSpot | Spot |
| JobTierBasic | Basic |
| JobTierStandard | Standard |
| JobTierPremium | Premium |
| Value |
|---|
Null |
Spot |
Basic |
Standard |
Premium |
| Name | Value |
|---|
| Null | Null |
| Spot | Spot |
| Basic | Basic |
| Standard | Standard |
| Premium | Premium |
| Name | Value |
|---|
| Null | Null |
| Spot | Spot |
| Basic | Basic |
| Standard | Standard |
| Premium | Premium |
| Name | Value |
|---|
| NULL | Null |
| SPOT | Spot |
| BASIC | Basic |
| STANDARD | Standard |
| PREMIUM | Premium |
| Value |
|---|
Null |
Spot |
Basic |
Standard |
Premium |
LearningRateScheduler
Learning rate scheduler enum.
| Name | Value | Description |
|---|
| None | None | No learning rate scheduler selected. |
| WarmupCosine | WarmupCosine | Cosine Annealing With Warmup. |
| Step | Step | Step learning rate scheduler. |
| Name | Value | Description |
|---|
| LearningRateSchedulerNone | None | No learning rate scheduler selected. |
| LearningRateSchedulerWarmupCosine | WarmupCosine | Cosine Annealing With Warmup. |
| LearningRateSchedulerStep | Step | Step learning rate scheduler. |
| Value | Description |
|---|
None | No learning rate scheduler selected. |
WarmupCosine | Cosine Annealing With Warmup. |
Step | Step learning rate scheduler. |
| Name | Value | Description |
|---|
| None | None | No learning rate scheduler selected. |
| WarmupCosine | WarmupCosine | Cosine Annealing With Warmup. |
| Step | Step | Step learning rate scheduler. |
| Name | Value | Description |
|---|
| None | None | No learning rate scheduler selected. |
| WarmupCosine | WarmupCosine | Cosine Annealing With Warmup. |
| Step | Step | Step learning rate scheduler. |
| Name | Value | Description |
|---|
| NONE | None | No learning rate scheduler selected. |
| WARMUP_COSINE | WarmupCosine | Cosine Annealing With Warmup. |
| STEP | Step | Step learning rate scheduler. |
| Value | Description |
|---|
None | No learning rate scheduler selected. |
WarmupCosine | Cosine Annealing With Warmup. |
Step | Step learning rate scheduler. |
Literal input type.
- Value
string
- [Required] Literal value for the input.
- Description
string
- Description for the input.
- Value
string
- [Required] Literal value for the input.
- Description
string
- Description for the input.
- value
string
- [Required] Literal value for the input.
- description
string
- Description for the input.
- value
String
- [Required] Literal value for the input.
- description
String
- Description for the input.
- value
string
- [Required] Literal value for the input.
- description
string
- Description for the input.
- value
str
- [Required] Literal value for the input.
- description
str
- Description for the input.
- value
String
- [Required] Literal value for the input.
- description
String
- Description for the input.
Literal input type.
- Value
string
- [Required] Literal value for the input.
- Description
string
- Description for the input.
- Value
string
- [Required] Literal value for the input.
- Description
string
- Description for the input.
- value
string
- [Required] Literal value for the input.
- description
string
- Description for the input.
- value
String
- [Required] Literal value for the input.
- description
String
- Description for the input.
- value
string
- [Required] Literal value for the input.
- description
string
- Description for the input.
- value
str
- [Required] Literal value for the input.
- description
str
- Description for the input.
- value
String
- [Required] Literal value for the input.
- description
String
- Description for the input.
LogVerbosity
Enum for setting log verbosity.
| Name | Value | Description |
|---|
| NotSet | NotSet | No logs emitted. |
| Debug | Debug | Debug and above log statements logged. |
| Info | Info | Info and above log statements logged. |
| Warning | Warning | Warning and above log statements logged. |
| Error | Error | Error and above log statements logged. |
| Critical | Critical | Only critical statements logged. |
| Name | Value | Description |
|---|
| LogVerbosityNotSet | NotSet | No logs emitted. |
| LogVerbosityDebug | Debug | Debug and above log statements logged. |
| LogVerbosityInfo | Info | Info and above log statements logged. |
| LogVerbosityWarning | Warning | Warning and above log statements logged. |
| LogVerbosityError | Error | Error and above log statements logged. |
| LogVerbosityCritical | Critical | Only critical statements logged. |
| Value | Description |
|---|
NotSet | No logs emitted. |
Debug | Debug and above log statements logged. |
Info | Info and above log statements logged. |
Warning | Warning and above log statements logged. |
Error | Error and above log statements logged. |
Critical | Only critical statements logged. |
| Name | Value | Description |
|---|
| NotSet | NotSet | No logs emitted. |
| Debug | Debug | Debug and above log statements logged. |
| Info | Info | Info and above log statements logged. |
| Warning | Warning | Warning and above log statements logged. |
| Error | Error | Error and above log statements logged. |
| Critical | Critical | Only critical statements logged. |
| Name | Value | Description |
|---|
| NotSet | NotSet | No logs emitted. |
| Debug | Debug | Debug and above log statements logged. |
| Info | Info | Info and above log statements logged. |
| Warning | Warning | Warning and above log statements logged. |
| Error | Error | Error and above log statements logged. |
| Critical | Critical | Only critical statements logged. |
| Name | Value | Description |
|---|
| NOT_SET | NotSet | No logs emitted. |
| DEBUG | Debug | Debug and above log statements logged. |
| INFO | Info | Info and above log statements logged. |
| WARNING | Warning | Warning and above log statements logged. |
| ERROR | Error | Error and above log statements logged. |
| CRITICAL | Critical | Only critical statements logged. |
| Value | Description |
|---|
NotSet | No logs emitted. |
Debug | Debug and above log statements logged. |
Info | Info and above log statements logged. |
Warning | Warning and above log statements logged. |
Error | Error and above log statements logged. |
Critical | Only critical statements logged. |
- Uri
string
- [Required] Input Asset URI.
- Description
string
- Description for the input.
- Mode
string
- Enum to determine the input data delivery mode.
- Uri
string
- [Required] Input Asset URI.
- Description
string
- Description for the input.
- Mode
string
- Enum to determine the input data delivery mode.
- uri
string
- [Required] Input Asset URI.
- description
string
- Description for the input.
- mode
string
- Enum to determine the input data delivery mode.
- uri
String
- [Required] Input Asset URI.
- description
String
- Description for the input.
- mode
String
- Enum to determine the input data delivery mode.
- uri
string
- [Required] Input Asset URI.
- description
string
- Description for the input.
- mode
string
- Enum to determine the input data delivery mode.
- uri
str
- [Required] Input Asset URI.
- description
str
- Description for the input.
- mode
str
- Enum to determine the input data delivery mode.
- uri
String
- [Required] Input Asset URI.
- description
String
- Description for the input.
- mode
String
- Enum to determine the input data delivery mode.
MLFlowModelJobOutput, MLFlowModelJobOutputArgs
MLFlowModelJobOutputResponse
- AssetName
string
- Output Asset Name.
- Description
string
- Description for the output.
- Mode
string
- Output data delivery mode enums.
- Uri
string
- Output Asset URI.
- AssetName
string
- Output Asset Name.
- Description
string
- Description for the output.
- Mode
string
- Output data delivery mode enums.
- Uri
string
- Output Asset URI.
- asset_name
string
- Output Asset Name.
- description
string
- Description for the output.
- mode
string
- Output data delivery mode enums.
- uri
string
- Output Asset URI.
- assetName
String
- Output Asset Name.
- description
String
- Description for the output.
- mode
String
- Output data delivery mode enums.
- uri
String
- Output Asset URI.
- assetName
string
- Output Asset Name.
- description
string
- Description for the output.
- mode
string
- Output data delivery mode enums.
- uri
string
- Output Asset URI.
- assetName
String
- Output Asset Name.
- description
String
- Description for the output.
- mode
String
- Output data delivery mode enums.
- uri
String
- Output Asset URI.
- Uri
string
- [Required] Input Asset URI.
- Description
string
- Description for the input.
- Mode
string
- Enum to determine the input data delivery mode.
- Uri
string
- [Required] Input Asset URI.
- Description
string
- Description for the input.
- Mode
string
- Enum to determine the input data delivery mode.
- uri
string
- [Required] Input Asset URI.
- description
string
- Description for the input.
- mode
string
- Enum to determine the input data delivery mode.
- uri
String
- [Required] Input Asset URI.
- description
String
- Description for the input.
- mode
String
- Enum to determine the input data delivery mode.
- uri
string
- [Required] Input Asset URI.
- description
string
- Description for the input.
- mode
string
- Enum to determine the input data delivery mode.
- uri
str
- [Required] Input Asset URI.
- description
str
- Description for the input.
- mode
str
- Enum to determine the input data delivery mode.
- uri
String
- [Required] Input Asset URI.
- description
String
- Description for the input.
- mode
String
- Enum to determine the input data delivery mode.
MLTableJobOutput, MLTableJobOutputArgs
MLTableJobOutputResponse
- AssetName
string
- Output Asset Name.
- Description
string
- Description for the output.
- Mode
string
- Output data delivery mode enums.
- Uri
string
- Output Asset URI.
- AssetName
string
- Output Asset Name.
- Description
string
- Description for the output.
- Mode
string
- Output data delivery mode enums.
- Uri
string
- Output Asset URI.
- asset_name
string
- Output Asset Name.
- description
string
- Description for the output.
- mode
string
- Output data delivery mode enums.
- uri
string
- Output Asset URI.
- assetName
String
- Output Asset Name.
- description
String
- Description for the output.
- mode
String
- Output data delivery mode enums.
- uri
String
- Output Asset URI.
- assetName
string
- Output Asset Name.
- description
string
- Description for the output.
- mode
string
- Output data delivery mode enums.
- uri
string
- Output Asset URI.
- assetName
String
- Output Asset Name.
- description
String
- Description for the output.
- mode
String
- Output data delivery mode enums.
- uri
String
- Output Asset URI.
ManagedComputeIdentity, ManagedComputeIdentityArgs
Managed compute identity definition.
- identity
object
- The identity which will be leveraged by the monitoring jobs.
ManagedComputeIdentityResponse
Managed compute identity definition.
- identity
object
- The identity which will be leveraged by the monitoring jobs.
ManagedIdentity, ManagedIdentityArgs
Managed identity configuration.
- ClientId
string
- Specifies a user-assigned identity by client ID. For system-assigned, do not set this field.
- 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.
- ClientId
string
- Specifies a user-assigned identity by client ID. For system-assigned, do not set this field.
- 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.
- client_id
string
- Specifies a user-assigned identity by client ID. For system-assigned, do not set this field.
- object_id
string
- Specifies a user-assigned identity by object ID. For system-assigned, do not set this field.
- resource_id
string
- Specifies a user-assigned identity by ARM resource ID. For system-assigned, do not set this field.
- clientId
String
- Specifies a user-assigned identity by client ID. For system-assigned, do not set this field.
- 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.
- clientId
string
- Specifies a user-assigned identity by client ID. For system-assigned, do not set this field.
- 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.
- client_id
str
- Specifies a user-assigned identity by client ID. For system-assigned, do not set this field.
- object_id
str
- Specifies a user-assigned identity by object ID. For system-assigned, do not set this field.
- resource_id
str
- Specifies a user-assigned identity by ARM resource ID. For system-assigned, do not set this field.
- clientId
String
- Specifies a user-assigned identity by client ID. For system-assigned, do not set this field.
- 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.
ManagedIdentityResponse
Managed identity configuration.
- ClientId
string
- Specifies a user-assigned identity by client ID. For system-assigned, do not set this field.
- 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.
- ClientId
string
- Specifies a user-assigned identity by client ID. For system-assigned, do not set this field.
- 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.
- client_id
string
- Specifies a user-assigned identity by client ID. For system-assigned, do not set this field.
- object_id
string
- Specifies a user-assigned identity by object ID. For system-assigned, do not set this field.
- resource_id
string
- Specifies a user-assigned identity by ARM resource ID. For system-assigned, do not set this field.
- clientId
String
- Specifies a user-assigned identity by client ID. For system-assigned, do not set this field.
- 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.
- clientId
string
- Specifies a user-assigned identity by client ID. For system-assigned, do not set this field.
- 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.
- client_id
str
- Specifies a user-assigned identity by client ID. For system-assigned, do not set this field.
- object_id
str
- Specifies a user-assigned identity by object ID. For system-assigned, do not set this field.
- resource_id
str
- Specifies a user-assigned identity by ARM resource ID. For system-assigned, do not set this field.
- clientId
String
- Specifies a user-assigned identity by client ID. For system-assigned, do not set this field.
- 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, ManagedServiceIdentityArgs
Managed service identity (system assigned and/or user assigned identities)
- Type
string | Pulumi.AzureNative.MachineLearningServices.ManagedServiceIdentityType
- Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed).
- UserAssignedIdentities
List<string>
- The set of user assigned identities associated with the resource. The userAssignedIdentities dictionary keys will be ARM resource ids in the form: '/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}. The dictionary values can be empty objects ({}) in requests.
- Type
string | ManagedServiceIdentityType
- Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed).
- UserAssignedIdentities
[]string
- The set of user assigned identities associated with the resource. The userAssignedIdentities dictionary keys will be ARM resource ids in the form: '/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}. The dictionary values can be empty objects ({}) in requests.
- type
string | "None" | "SystemAssigned" | "UserAssigned" | "SystemAssigned,UserAssigned"
- Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed).
- user_assigned_identities
list(string)
- The set of user assigned identities associated with the resource. The userAssignedIdentities dictionary keys will be ARM resource ids in the form: '/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}. The dictionary values can be empty objects ({}) in requests.
- type
String | ManagedServiceIdentityType
- Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed).
- userAssignedIdentities
List<String>
- The set of user assigned identities associated with the resource. The userAssignedIdentities dictionary keys will be ARM resource ids in the form: '/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}. The dictionary values can be empty objects ({}) in requests.
- type
string | ManagedServiceIdentityType
- Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed).
- userAssignedIdentities
string[]
- The set of user assigned identities associated with the resource. The userAssignedIdentities dictionary keys will be ARM resource ids in the form: '/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}. The dictionary values can be empty objects ({}) in requests.
- type
str | ManagedServiceIdentityType
- Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed).
- user_assigned_identities
Sequence[str]
- The set of user assigned identities associated with the resource. The userAssignedIdentities dictionary keys will be ARM resource ids in the form: '/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}. The dictionary values can be empty objects ({}) in requests.
- type
String | "None" | "SystemAssigned" | "UserAssigned" | "SystemAssigned,UserAssigned"
- Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed).
- userAssignedIdentities
List<String>
- The set of user assigned identities associated with the resource. The userAssignedIdentities dictionary keys will be ARM resource ids in the form: '/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}. The dictionary values can be empty objects ({}) in requests.
ManagedServiceIdentityResponse
Managed service identity (system assigned and/or user assigned identities)
- 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
string
- Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed).
- UserAssignedIdentities
Dictionary<string, Pulumi.AzureNative.MachineLearningServices.Inputs.UserAssignedIdentityResponse>
- The set of user assigned identities associated with the resource. The userAssignedIdentities dictionary keys will be ARM resource ids in the form: '/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}. The dictionary values can be empty objects ({}) in requests.
- 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
string
- Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed).
- UserAssignedIdentities
map[string]UserAssignedIdentityResponse
- The set of user assigned identities associated with the resource. The userAssignedIdentities dictionary keys will be ARM resource ids in the form: '/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}. The dictionary values can be empty objects ({}) in requests.
- principal_id
string
- The service principal ID of the system assigned identity. This property will only be provided for a system assigned identity.
- tenant_id
string
- The tenant ID of the system assigned identity. This property will only be provided for a system assigned identity.
- type
string
- Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed).
- user_assigned_identities
map(object)
- The set of user assigned identities associated with the resource. The userAssignedIdentities dictionary keys will be ARM resource ids in the form: '/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}. The dictionary values can be empty objects ({}) in requests.
- 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
String
- Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed).
- userAssignedIdentities
Map<String,UserAssignedIdentityResponse>
- The set of user assigned identities associated with the resource. The userAssignedIdentities dictionary keys will be ARM resource ids in the form: '/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}. The dictionary values can be empty objects ({}) in requests.
- 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
string
- Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed).
- userAssignedIdentities
{[key: string]: UserAssignedIdentityResponse}
- The set of user assigned identities associated with the resource. The userAssignedIdentities dictionary keys will be ARM resource ids in the form: '/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}. The dictionary values can be empty objects ({}) in requests.
- principal_id
str
- The service principal ID of the system assigned identity. This property will only be provided for a system assigned identity.
- tenant_id
str
- The tenant ID of the system assigned identity. This property will only be provided for a system assigned identity.
- type
str
- Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed).
- user_assigned_identities
Mapping[str, UserAssignedIdentityResponse]
- The set of user assigned identities associated with the resource. The userAssignedIdentities dictionary keys will be ARM resource ids in the form: '/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}. The dictionary values can be empty objects ({}) in requests.
- 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
String
- Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed).
- userAssignedIdentities
Map<Property Map>
- The set of user assigned identities associated with the resource. The userAssignedIdentities dictionary keys will be ARM resource ids in the form: '/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}. The dictionary values can be empty objects ({}) in requests.
ManagedServiceIdentityType
Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed).
| Name | Value |
|---|
| None | None |
| SystemAssigned | SystemAssigned |
| UserAssigned | UserAssigned |
| SystemAssigned_UserAssigned | SystemAssigned,UserAssigned |
| Name | Value |
|---|
| ManagedServiceIdentityTypeNone | None |
| ManagedServiceIdentityTypeSystemAssigned | SystemAssigned |
| ManagedServiceIdentityTypeUserAssigned | UserAssigned |
| ManagedServiceIdentityType_SystemAssigned_UserAssigned | SystemAssigned,UserAssigned |
| Value |
|---|
None |
SystemAssigned |
UserAssigned |
SystemAssigned,UserAssigned |
| Name | Value |
|---|
| None | None |
| SystemAssigned | SystemAssigned |
| UserAssigned | UserAssigned |
| SystemAssigned_UserAssigned | SystemAssigned,UserAssigned |
| Name | Value |
|---|
| None | None |
| SystemAssigned | SystemAssigned |
| UserAssigned | UserAssigned |
| SystemAssigned_UserAssigned | SystemAssigned,UserAssigned |
| Name | Value |
|---|
| NONE | None |
| SYSTEM_ASSIGNED | SystemAssigned |
| USER_ASSIGNED | UserAssigned |
| SYSTEM_ASSIGNED_USER_ASSIGNED | SystemAssigned,UserAssigned |
| Value |
|---|
None |
SystemAssigned |
UserAssigned |
SystemAssigned,UserAssigned |
Defines an early termination policy based on running averages of the primary metric of all runs
Defines an early termination policy based on running averages of the primary metric of all runs
ModelSize
Image model size.
| Name | Value | Description |
|---|
| None | None | No value selected. |
| Small | Small | Small size. |
| Medium | Medium | Medium size. |
| Large | Large | Large size. |
| ExtraLarge | ExtraLarge | Extra large size. |
| Name | Value | Description |
|---|
| ModelSizeNone | None | No value selected. |
| ModelSizeSmall | Small | Small size. |
| ModelSizeMedium | Medium | Medium size. |
| ModelSizeLarge | Large | Large size. |
| ModelSizeExtraLarge | ExtraLarge | Extra large size. |
| Value | Description |
|---|
None | No value selected. |
Small | Small size. |
Medium | Medium size. |
Large | Large size. |
ExtraLarge | Extra large size. |
| Name | Value | Description |
|---|
| None | None | No value selected. |
| Small | Small | Small size. |
| Medium | Medium | Medium size. |
| Large | Large | Large size. |
| ExtraLarge | ExtraLarge | Extra large size. |
| Name | Value | Description |
|---|
| None | None | No value selected. |
| Small | Small | Small size. |
| Medium | Medium | Medium size. |
| Large | Large | Large size. |
| ExtraLarge | ExtraLarge | Extra large size. |
| Name | Value | Description |
|---|
| NONE | None | No value selected. |
| SMALL | Small | Small size. |
| MEDIUM | Medium | Medium size. |
| LARGE | Large | Large size. |
| EXTRA_LARGE | ExtraLarge | Extra large size. |
| Value | Description |
|---|
None | No value selected. |
Small | Small size. |
Medium | Medium size. |
Large | Large size. |
ExtraLarge | Extra large size. |
ModelTaskType
[Required] The machine learning task type of the monitored model.
| Name | Value |
|---|
| Classification | Classification |
| Regression | Regression |
| Name | Value |
|---|
| ModelTaskTypeClassification | Classification |
| ModelTaskTypeRegression | Regression |
| Value |
|---|
Classification |
Regression |
| Name | Value |
|---|
| Classification | Classification |
| Regression | Regression |
| Name | Value |
|---|
| Classification | Classification |
| Regression | Regression |
| Name | Value |
|---|
| CLASSIFICATION | Classification |
| REGRESSION | Regression |
| Value |
|---|
Classification |
Regression |
MonitorDefinition, MonitorDefinitionArgs
MonitorDefinitionResponse
MonitorEmailNotificationSettings, MonitorEmailNotificationSettingsArgs
- Emails
List<string>
- The email recipient list which has a limitation of 499 characters in total.
- Emails
[]string
- The email recipient list which has a limitation of 499 characters in total.
- emails
list(string)
- The email recipient list which has a limitation of 499 characters in total.
- emails
List<String>
- The email recipient list which has a limitation of 499 characters in total.
- emails
string[]
- The email recipient list which has a limitation of 499 characters in total.
- emails
Sequence[str]
- The email recipient list which has a limitation of 499 characters in total.
- emails
List<String>
- The email recipient list which has a limitation of 499 characters in total.
MonitorEmailNotificationSettingsResponse
- Emails
List<string>
- The email recipient list which has a limitation of 499 characters in total.
- Emails
[]string
- The email recipient list which has a limitation of 499 characters in total.
- emails
list(string)
- The email recipient list which has a limitation of 499 characters in total.
- emails
List<String>
- The email recipient list which has a limitation of 499 characters in total.
- emails
string[]
- The email recipient list which has a limitation of 499 characters in total.
- emails
Sequence[str]
- The email recipient list which has a limitation of 499 characters in total.
- emails
List<String>
- The email recipient list which has a limitation of 499 characters in total.
MonitorNotificationSettings, MonitorNotificationSettingsArgs
MonitorNotificationSettingsResponse
MonitorServerlessSparkCompute, MonitorServerlessSparkComputeArgs
Monitor serverless spark compute definition.
MonitorServerlessSparkComputeResponse
Monitor serverless spark compute definition.
MonitoringFeatureDataType
[Required] Specifies the data type of the metric threshold.
| Name | Value | Description |
|---|
| Numerical | Numerical | Used for features of numerical data type. |
| Categorical | Categorical | Used for features of categorical data type. |
| Name | Value | Description |
|---|
| MonitoringFeatureDataTypeNumerical | Numerical | Used for features of numerical data type. |
| MonitoringFeatureDataTypeCategorical | Categorical | Used for features of categorical data type. |
| Value | Description |
|---|
Numerical | Used for features of numerical data type. |
Categorical | Used for features of categorical data type. |
| Name | Value | Description |
|---|
| Numerical | Numerical | Used for features of numerical data type. |
| Categorical | Categorical | Used for features of categorical data type. |
| Name | Value | Description |
|---|
| Numerical | Numerical | Used for features of numerical data type. |
| Categorical | Categorical | Used for features of categorical data type. |
| Name | Value | Description |
|---|
| NUMERICAL | Numerical | Used for features of numerical data type. |
| CATEGORICAL | Categorical | Used for features of categorical data type. |
| Value | Description |
|---|
Numerical | Used for features of numerical data type. |
Categorical | Used for features of categorical data type. |
MonitoringNotificationType
| Name | Value | Description |
|---|
| AmlNotification | AmlNotification | Enables email notifications through AML notifications. |
| Name | Value | Description |
|---|
| MonitoringNotificationTypeAmlNotification | AmlNotification | Enables email notifications through AML notifications. |
| Value | Description |
|---|
AmlNotification | Enables email notifications through AML notifications. |
| Name | Value | Description |
|---|
| AmlNotification | AmlNotification | Enables email notifications through AML notifications. |
| Name | Value | Description |
|---|
| AmlNotification | AmlNotification | Enables email notifications through AML notifications. |
| Name | Value | Description |
|---|
| AML_NOTIFICATION | AmlNotification | Enables email notifications through AML notifications. |
| Value | Description |
|---|
AmlNotification | Enables email notifications through AML notifications. |
MonitoringTarget, MonitoringTargetArgs
Monitoring target definition.
- TaskType
string | ModelTaskType
- [Required] The machine learning task type of the monitored model.
- DeploymentId
string
- Reference to the deployment asset targeted by this monitor.
- ModelId
string
- Reference to the model asset targeted by this monitor.
- taskType
String | ModelTaskType
- [Required] The machine learning task type of the monitored model.
- deploymentId
String
- Reference to the deployment asset targeted by this monitor.
- modelId
String
- Reference to the model asset targeted by this monitor.
- taskType
string | ModelTaskType
- [Required] The machine learning task type of the monitored model.
- deploymentId
string
- Reference to the deployment asset targeted by this monitor.
- modelId
string
- Reference to the model asset targeted by this monitor.
- task_type
str | ModelTaskType
- [Required] The machine learning task type of the monitored model.
- deployment_id
str
- Reference to the deployment asset targeted by this monitor.
- model_id
str
- Reference to the model asset targeted by this monitor.
MonitoringTargetResponse
Monitoring target definition.
- TaskType
string
- [Required] The machine learning task type of the monitored model.
- DeploymentId
string
- Reference to the deployment asset targeted by this monitor.
- ModelId
string
- Reference to the model asset targeted by this monitor.
- TaskType
string
- [Required] The machine learning task type of the monitored model.
- DeploymentId
string
- Reference to the deployment asset targeted by this monitor.
- ModelId
string
- Reference to the model asset targeted by this monitor.
- task_type
string
- [Required] The machine learning task type of the monitored model.
- deployment_id
string
- Reference to the deployment asset targeted by this monitor.
- model_id
string
- Reference to the model asset targeted by this monitor.
- taskType
String
- [Required] The machine learning task type of the monitored model.
- deploymentId
String
- Reference to the deployment asset targeted by this monitor.
- modelId
String
- Reference to the model asset targeted by this monitor.
- taskType
string
- [Required] The machine learning task type of the monitored model.
- deploymentId
string
- Reference to the deployment asset targeted by this monitor.
- modelId
string
- Reference to the model asset targeted by this monitor.
- task_type
str
- [Required] The machine learning task type of the monitored model.
- deployment_id
str
- Reference to the deployment asset targeted by this monitor.
- model_id
str
- Reference to the model asset targeted by this monitor.
- taskType
String
- [Required] The machine learning task type of the monitored model.
- deploymentId
String
- Reference to the deployment asset targeted by this monitor.
- modelId
String
- Reference to the model asset targeted by this monitor.
MonitoringThreshold, MonitoringThresholdArgs
- Value
double
- The threshold value. If null, the set default is dependent on the metric type.
- Value
float64
- The threshold value. If null, the set default is dependent on the metric type.
- value
number
- The threshold value. If null, the set default is dependent on the metric type.
- value
Double
- The threshold value. If null, the set default is dependent on the metric type.
- value
number
- The threshold value. If null, the set default is dependent on the metric type.
- value
float
- The threshold value. If null, the set default is dependent on the metric type.
- value
Number
- The threshold value. If null, the set default is dependent on the metric type.
MonitoringThresholdResponse
- Value
double
- The threshold value. If null, the set default is dependent on the metric type.
- Value
float64
- The threshold value. If null, the set default is dependent on the metric type.
- value
number
- The threshold value. If null, the set default is dependent on the metric type.
- value
Double
- The threshold value. If null, the set default is dependent on the metric type.
- value
number
- The threshold value. If null, the set default is dependent on the metric type.
- value
float
- The threshold value. If null, the set default is dependent on the metric type.
- value
Number
- The threshold value. If null, the set default is dependent on the metric type.
Mpi, MpiArgs
MPI distribution configuration.
MpiResponse
MPI distribution configuration.
NlpVerticalFeaturizationSettings, NlpVerticalFeaturizationSettingsArgs
NlpVerticalFeaturizationSettingsResponse
NlpVerticalLimitSettings, NlpVerticalLimitSettingsArgs
Job execution constraints.
NlpVerticalLimitSettingsResponse
Job execution constraints.
NotificationSetting, NotificationSettingArgs
Configuration for notification.
- EmailOn
List<Union<string, Pulumi.AzureNative.MachineLearningServices.EmailNotificationEnableType>>
- Send email notification to user on specified notification type
- Emails
List<string>
- This is the email recipient list which has a limitation of 499 characters in total concat with comma separator
- Webhooks
Dictionary<string, Pulumi.AzureNative.MachineLearningServices.Inputs.AzureDevOpsWebhook>
- Send webhook callback to a service. Key is a user-provided name for the webhook.
- EmailOn
[]string
- Send email notification to user on specified notification type
- Emails
[]string
- This is the email recipient list which has a limitation of 499 characters in total concat with comma separator
- Webhooks
map[string]AzureDevOpsWebhook
- Send webhook callback to a service. Key is a user-provided name for the webhook.
- email_on
list(string | "JobCompleted" | "JobFailed" | "JobCancelled")
- Send email notification to user on specified notification type
- emails
list(string)
- This is the email recipient list which has a limitation of 499 characters in total concat with comma separator
- webhooks
map(object)
- Send webhook callback to a service. Key is a user-provided name for the webhook.
- emailOn
List<Either<String,EmailNotificationEnableType>>
- Send email notification to user on specified notification type
- emails
List<String>
- This is the email recipient list which has a limitation of 499 characters in total concat with comma separator
- webhooks
Map<String,AzureDevOpsWebhook>
- Send webhook callback to a service. Key is a user-provided name for the webhook.
- emailOn
(string | EmailNotificationEnableType)[]
- Send email notification to user on specified notification type
- emails
string[]
- This is the email recipient list which has a limitation of 499 characters in total concat with comma separator
- webhooks
{[key: string]: AzureDevOpsWebhook}
- Send webhook callback to a service. Key is a user-provided name for the webhook.
- email_on
Sequence[Union[str, EmailNotificationEnableType]]
- Send email notification to user on specified notification type
- emails
Sequence[str]
- This is the email recipient list which has a limitation of 499 characters in total concat with comma separator
- webhooks
Mapping[str, AzureDevOpsWebhook]
- Send webhook callback to a service. Key is a user-provided name for the webhook.
- emailOn
List<String | "JobCompleted" | "JobFailed" | "JobCancelled">
- Send email notification to user on specified notification type
- emails
List<String>
- This is the email recipient list which has a limitation of 499 characters in total concat with comma separator
- webhooks
Map<Property Map>
- Send webhook callback to a service. Key is a user-provided name for the webhook.
NotificationSettingResponse
Configuration for notification.
- EmailOn
List<string>
- Send email notification to user on specified notification type
- Emails
List<string>
- This is the email recipient list which has a limitation of 499 characters in total concat with comma separator
- Webhooks
Dictionary<string, Pulumi.AzureNative.MachineLearningServices.Inputs.AzureDevOpsWebhookResponse>
- Send webhook callback to a service. Key is a user-provided name for the webhook.
- EmailOn
[]string
- Send email notification to user on specified notification type
- Emails
[]string
- This is the email recipient list which has a limitation of 499 characters in total concat with comma separator
- Webhooks
map[string]AzureDevOpsWebhookResponse
- Send webhook callback to a service. Key is a user-provided name for the webhook.
- email_on
list(string)
- Send email notification to user on specified notification type
- emails
list(string)
- This is the email recipient list which has a limitation of 499 characters in total concat with comma separator
- webhooks
map(object)
- Send webhook callback to a service. Key is a user-provided name for the webhook.
- emailOn
List<String>
- Send email notification to user on specified notification type
- emails
List<String>
- This is the email recipient list which has a limitation of 499 characters in total concat with comma separator
- webhooks
Map<String,AzureDevOpsWebhookResponse>
- Send webhook callback to a service. Key is a user-provided name for the webhook.
- emailOn
string[]
- Send email notification to user on specified notification type
- emails
string[]
- This is the email recipient list which has a limitation of 499 characters in total concat with comma separator
- webhooks
{[key: string]: AzureDevOpsWebhookResponse}
- Send webhook callback to a service. Key is a user-provided name for the webhook.
- email_on
Sequence[str]
- Send email notification to user on specified notification type
- emails
Sequence[str]
- This is the email recipient list which has a limitation of 499 characters in total concat with comma separator
- webhooks
Mapping[str, AzureDevOpsWebhookResponse]
- Send webhook callback to a service. Key is a user-provided name for the webhook.
- emailOn
List<String>
- Send email notification to user on specified notification type
- emails
List<String>
- This is the email recipient list which has a limitation of 499 characters in total concat with comma separator
- webhooks
Map<Property Map>
- Send webhook callback to a service. Key is a user-provided name for the webhook.
NumericalDataDriftMetric
[Required] The numerical data drift metric to calculate.
| Name | Value | Description |
|---|
| JensenShannonDistance | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| PopulationStabilityIndex | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| NormalizedWassersteinDistance | NormalizedWassersteinDistance | The Normalized Wasserstein Distance metric. |
| TwoSampleKolmogorovSmirnovTest | TwoSampleKolmogorovSmirnovTest | The Two Sample Kolmogorov-Smirnov Test (two-sample K–S) metric. |
| Name | Value | Description |
|---|
| NumericalDataDriftMetricJensenShannonDistance | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| NumericalDataDriftMetricPopulationStabilityIndex | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| NumericalDataDriftMetricNormalizedWassersteinDistance | NormalizedWassersteinDistance | The Normalized Wasserstein Distance metric. |
| NumericalDataDriftMetricTwoSampleKolmogorovSmirnovTest | TwoSampleKolmogorovSmirnovTest | The Two Sample Kolmogorov-Smirnov Test (two-sample K–S) metric. |
| Value | Description |
|---|
JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
NormalizedWassersteinDistance | The Normalized Wasserstein Distance metric. |
TwoSampleKolmogorovSmirnovTest | The Two Sample Kolmogorov-Smirnov Test (two-sample K–S) metric. |
| Name | Value | Description |
|---|
| JensenShannonDistance | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| PopulationStabilityIndex | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| NormalizedWassersteinDistance | NormalizedWassersteinDistance | The Normalized Wasserstein Distance metric. |
| TwoSampleKolmogorovSmirnovTest | TwoSampleKolmogorovSmirnovTest | The Two Sample Kolmogorov-Smirnov Test (two-sample K–S) metric. |
| Name | Value | Description |
|---|
| JensenShannonDistance | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| PopulationStabilityIndex | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| NormalizedWassersteinDistance | NormalizedWassersteinDistance | The Normalized Wasserstein Distance metric. |
| TwoSampleKolmogorovSmirnovTest | TwoSampleKolmogorovSmirnovTest | The Two Sample Kolmogorov-Smirnov Test (two-sample K–S) metric. |
| Name | Value | Description |
|---|
| JENSEN_SHANNON_DISTANCE | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| POPULATION_STABILITY_INDEX | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| NORMALIZED_WASSERSTEIN_DISTANCE | NormalizedWassersteinDistance | The Normalized Wasserstein Distance metric. |
| TWO_SAMPLE_KOLMOGOROV_SMIRNOV_TEST | TwoSampleKolmogorovSmirnovTest | The Two Sample Kolmogorov-Smirnov Test (two-sample K–S) metric. |
| Value | Description |
|---|
JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
NormalizedWassersteinDistance | The Normalized Wasserstein Distance metric. |
TwoSampleKolmogorovSmirnovTest | The Two Sample Kolmogorov-Smirnov Test (two-sample K–S) metric. |
NumericalDataDriftMetricThreshold, NumericalDataDriftMetricThresholdArgs
NumericalDataDriftMetricThresholdResponse
- metric
string
- [Required] The numerical data drift metric to calculate.
- threshold
object
- The threshold value. If null, a default value will be set depending on the selected metric.
- metric
String
- [Required] The numerical data drift metric to calculate.
- threshold
Property Map
- The threshold value. If null, a default value will be set depending on the selected metric.
NumericalDataQualityMetric
[Required] The numerical data quality metric to calculate.
| Name | Value | Description |
|---|
| NullValueRate | NullValueRate | Calculates the rate of null values. |
| DataTypeErrorRate | DataTypeErrorRate | Calculates the rate of data type errors. |
| OutOfBoundsRate | OutOfBoundsRate | Calculates the rate values are out of bounds. |
| Name | Value | Description |
|---|
| NumericalDataQualityMetricNullValueRate | NullValueRate | Calculates the rate of null values. |
| NumericalDataQualityMetricDataTypeErrorRate | DataTypeErrorRate | Calculates the rate of data type errors. |
| NumericalDataQualityMetricOutOfBoundsRate | OutOfBoundsRate | Calculates the rate values are out of bounds. |
| Value | Description |
|---|
NullValueRate | Calculates the rate of null values. |
DataTypeErrorRate | Calculates the rate of data type errors. |
OutOfBoundsRate | Calculates the rate values are out of bounds. |
| Name | Value | Description |
|---|
| NullValueRate | NullValueRate | Calculates the rate of null values. |
| DataTypeErrorRate | DataTypeErrorRate | Calculates the rate of data type errors. |
| OutOfBoundsRate | OutOfBoundsRate | Calculates the rate values are out of bounds. |
| Name | Value | Description |
|---|
| NullValueRate | NullValueRate | Calculates the rate of null values. |
| DataTypeErrorRate | DataTypeErrorRate | Calculates the rate of data type errors. |
| OutOfBoundsRate | OutOfBoundsRate | Calculates the rate values are out of bounds. |
| Name | Value | Description |
|---|
| NULL_VALUE_RATE | NullValueRate | Calculates the rate of null values. |
| DATA_TYPE_ERROR_RATE | DataTypeErrorRate | Calculates the rate of data type errors. |
| OUT_OF_BOUNDS_RATE | OutOfBoundsRate | Calculates the rate values are out of bounds. |
| Value | Description |
|---|
NullValueRate | Calculates the rate of null values. |
DataTypeErrorRate | Calculates the rate of data type errors. |
OutOfBoundsRate | Calculates the rate values are out of bounds. |
NumericalDataQualityMetricThreshold, NumericalDataQualityMetricThresholdArgs
NumericalDataQualityMetricThresholdResponse
- metric
string
- [Required] The numerical data quality metric to calculate.
- threshold
object
- The threshold value. If null, a default value will be set depending on the selected metric.
- metric
String
- [Required] The numerical data quality metric to calculate.
- threshold
Property Map
- The threshold value. If null, a default value will be set depending on the selected metric.
NumericalPredictionDriftMetric
[Required] The numerical prediction drift metric to calculate.
| Name | Value | Description |
|---|
| JensenShannonDistance | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| PopulationStabilityIndex | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| NormalizedWassersteinDistance | NormalizedWassersteinDistance | The Normalized Wasserstein Distance metric. |
| TwoSampleKolmogorovSmirnovTest | TwoSampleKolmogorovSmirnovTest | The Two Sample Kolmogorov-Smirnov Test (two-sample K–S) metric. |
| Name | Value | Description |
|---|
| NumericalPredictionDriftMetricJensenShannonDistance | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| NumericalPredictionDriftMetricPopulationStabilityIndex | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| NumericalPredictionDriftMetricNormalizedWassersteinDistance | NormalizedWassersteinDistance | The Normalized Wasserstein Distance metric. |
| NumericalPredictionDriftMetricTwoSampleKolmogorovSmirnovTest | TwoSampleKolmogorovSmirnovTest | The Two Sample Kolmogorov-Smirnov Test (two-sample K–S) metric. |
| Value | Description |
|---|
JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
NormalizedWassersteinDistance | The Normalized Wasserstein Distance metric. |
TwoSampleKolmogorovSmirnovTest | The Two Sample Kolmogorov-Smirnov Test (two-sample K–S) metric. |
| Name | Value | Description |
|---|
| JensenShannonDistance | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| PopulationStabilityIndex | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| NormalizedWassersteinDistance | NormalizedWassersteinDistance | The Normalized Wasserstein Distance metric. |
| TwoSampleKolmogorovSmirnovTest | TwoSampleKolmogorovSmirnovTest | The Two Sample Kolmogorov-Smirnov Test (two-sample K–S) metric. |
| Name | Value | Description |
|---|
| JensenShannonDistance | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| PopulationStabilityIndex | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| NormalizedWassersteinDistance | NormalizedWassersteinDistance | The Normalized Wasserstein Distance metric. |
| TwoSampleKolmogorovSmirnovTest | TwoSampleKolmogorovSmirnovTest | The Two Sample Kolmogorov-Smirnov Test (two-sample K–S) metric. |
| Name | Value | Description |
|---|
| JENSEN_SHANNON_DISTANCE | JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
| POPULATION_STABILITY_INDEX | PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
| NORMALIZED_WASSERSTEIN_DISTANCE | NormalizedWassersteinDistance | The Normalized Wasserstein Distance metric. |
| TWO_SAMPLE_KOLMOGOROV_SMIRNOV_TEST | TwoSampleKolmogorovSmirnovTest | The Two Sample Kolmogorov-Smirnov Test (two-sample K–S) metric. |
| Value | Description |
|---|
JensenShannonDistance | The Jensen Shannon Distance (JSD) metric. |
PopulationStabilityIndex | The Population Stability Index (PSI) metric. |
NormalizedWassersteinDistance | The Normalized Wasserstein Distance metric. |
TwoSampleKolmogorovSmirnovTest | The Two Sample Kolmogorov-Smirnov Test (two-sample K–S) metric. |
NumericalPredictionDriftMetricThreshold, NumericalPredictionDriftMetricThresholdArgs
NumericalPredictionDriftMetricThresholdResponse
- metric
string
- [Required] The numerical prediction drift metric to calculate.
- threshold
object
- The threshold value. If null, a default value will be set depending on the selected metric.
- metric
String
- [Required] The numerical prediction drift metric to calculate.
- threshold
Property Map
- The threshold value. If null, a default value will be set depending on the selected metric.
ObjectDetectionPrimaryMetrics
Primary metrics for Image ObjectDetection task.
| Name | Value | Description |
|---|
| MeanAveragePrecision | MeanAveragePrecision | Mean Average Precision (MAP) is the average of AP (Average Precision).
AP is calculated for each class and averaged to get the MAP. |
| Name | Value | Description |
|---|
| ObjectDetectionPrimaryMetricsMeanAveragePrecision | MeanAveragePrecision | Mean Average Precision (MAP) is the average of AP (Average Precision).
AP is calculated for each class and averaged to get the MAP. |
| Value | Description |
|---|
MeanAveragePrecision | Mean Average Precision (MAP) is the average of AP (Average Precision).
AP is calculated for each class and averaged to get the MAP. |
| Name | Value | Description |
|---|
| MeanAveragePrecision | MeanAveragePrecision | Mean Average Precision (MAP) is the average of AP (Average Precision).
AP is calculated for each class and averaged to get the MAP. |
| Name | Value | Description |
|---|
| MeanAveragePrecision | MeanAveragePrecision | Mean Average Precision (MAP) is the average of AP (Average Precision).
AP is calculated for each class and averaged to get the MAP. |
| Name | Value | Description |
|---|
| MEAN_AVERAGE_PRECISION | MeanAveragePrecision | Mean Average Precision (MAP) is the average of AP (Average Precision).
AP is calculated for each class and averaged to get the MAP. |
| Value | Description |
|---|
MeanAveragePrecision | Mean Average Precision (MAP) is the average of AP (Average Precision).
AP is calculated for each class and averaged to get the MAP. |
Objective, ObjectiveArgs
Optimization objective.
- Goal
string | Goal
- [Required] Defines supported metric goals for hyperparameter tuning
- PrimaryMetric
string
- [Required] Name of the metric to optimize.
- goal
String | Goal
- [Required] Defines supported metric goals for hyperparameter tuning
- primaryMetric
String
- [Required] Name of the metric to optimize.
- goal
string | Goal
- [Required] Defines supported metric goals for hyperparameter tuning
- primaryMetric
string
- [Required] Name of the metric to optimize.
- goal
str | Goal
- [Required] Defines supported metric goals for hyperparameter tuning
- primary_metric
str
- [Required] Name of the metric to optimize.
ObjectiveResponse
Optimization objective.
- Goal
string
- [Required] Defines supported metric goals for hyperparameter tuning
- PrimaryMetric
string
- [Required] Name of the metric to optimize.
- Goal
string
- [Required] Defines supported metric goals for hyperparameter tuning
- PrimaryMetric
string
- [Required] Name of the metric to optimize.
- goal
string
- [Required] Defines supported metric goals for hyperparameter tuning
- primary_metric
string
- [Required] Name of the metric to optimize.
- goal
String
- [Required] Defines supported metric goals for hyperparameter tuning
- primaryMetric
String
- [Required] Name of the metric to optimize.
- goal
string
- [Required] Defines supported metric goals for hyperparameter tuning
- primaryMetric
string
- [Required] Name of the metric to optimize.
- goal
str
- [Required] Defines supported metric goals for hyperparameter tuning
- primary_metric
str
- [Required] Name of the metric to optimize.
- goal
String
- [Required] Defines supported metric goals for hyperparameter tuning
- primaryMetric
String
- [Required] Name of the metric to optimize.
OutputDeliveryMode
Output data delivery mode enums.
| Name | Value |
|---|
| ReadWriteMount | ReadWriteMount |
| Upload | Upload |
| Direct | Direct |
| Name | Value |
|---|
| OutputDeliveryModeReadWriteMount | ReadWriteMount |
| OutputDeliveryModeUpload | Upload |
| OutputDeliveryModeDirect | Direct |
| Value |
|---|
ReadWriteMount |
Upload |
Direct |
| Name | Value |
|---|
| ReadWriteMount | ReadWriteMount |
| Upload | Upload |
| Direct | Direct |
| Name | Value |
|---|
| ReadWriteMount | ReadWriteMount |
| Upload | Upload |
| Direct | Direct |
| Name | Value |
|---|
| READ_WRITE_MOUNT | ReadWriteMount |
| UPLOAD | Upload |
| DIRECT | Direct |
| Value |
|---|
ReadWriteMount |
Upload |
Direct |
PipelineJob, PipelineJobArgs
Pipeline Job definition: defines generic to MFE attributes.
- 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
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- Identity
Pulumi.AzureNative.MachineLearningServices.Inputs.AmlToken | Pulumi.AzureNative.MachineLearningServices.Inputs.ManagedIdentity | Pulumi.AzureNative.MachineLearningServices.Inputs.UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- Inputs
Dictionary<string, object>
- Inputs for the pipeline job.
- IsArchived
bool
- Is the asset archived?
- Jobs
Dictionary<string, object>
- Jobs construct the Pipeline Job.
- NotificationSetting
Pulumi.AzureNative.MachineLearningServices.Inputs.NotificationSetting
- Notification setting for the job
- Outputs
Dictionary<string, object>
- Outputs for the pipeline job
- Properties
Dictionary<string, string>
- The asset property dictionary.
- Services
Dictionary<string, Pulumi.AzureNative.MachineLearningServices.Inputs.JobService>
- 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.
- Tags
Dictionary<string, string>
- Tag dictionary. Tags can be added, removed, and updated.
- 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
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- Identity
AmlToken | ManagedIdentity | UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- Inputs
map[string]interface{}
- Inputs for the pipeline job.
- IsArchived
bool
- Is the asset archived?
- Jobs
map[string]interface{}
- Jobs construct the Pipeline Job.
- NotificationSetting
NotificationSetting
- Notification setting for the job
- Outputs
map[string]interface{}
- Outputs for the pipeline job
- Properties
map[string]string
- The asset property dictionary.
- Services
map[string]JobService
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- Settings
interface{}
- Pipeline settings, for things like ContinueRunOnStepFailure etc.
- SourceJobId
string
- ARM resource ID of source job.
- Tags
map[string]string
- Tag dictionary. Tags can be added, removed, and updated.
- component_id
string
- ARM resource ID of the component resource.
- compute_id
string
- ARM resource ID of the compute resource.
- description
string
- The asset description text.
- display_name
string
- Display name of job.
- experiment_name
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
object | object | object
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
map(object | object | object | object | object | object | object)
- Inputs for the pipeline job.
- is_archived
bool
- Is the asset archived?
- jobs
map(any)
- Jobs construct the Pipeline Job.
- notification_setting
object
- Notification setting for the job
- outputs
map(object | object | object | object | object | object)
- Outputs for the pipeline job
- properties
map(string)
- The asset property dictionary.
- services
map(object)
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- settings
any
- Pipeline settings, for things like ContinueRunOnStepFailure etc.
- source_job_id
string
- ARM resource ID of source job.
- tags
map(string)
- Tag dictionary. Tags can be added, removed, and updated.
- 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
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlToken | ManagedIdentity | UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Map<String,Object>
- Inputs for the pipeline job.
- isArchived
Boolean
- Is the asset archived?
- jobs
Map<String,Object>
- Jobs construct the Pipeline Job.
- notificationSetting
NotificationSetting
- Notification setting for the job
- outputs
Map<String,Object>
- Outputs for the pipeline job
- properties
Map<String,String>
- The asset property dictionary.
- services
Map<String,JobService>
- 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.
- tags
Map<String,String>
- Tag dictionary. Tags can be added, removed, and updated.
- 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
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlToken | ManagedIdentity | UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
{[key: string]: CustomModelJobInput | LiteralJobInput | MLFlowModelJobInput | MLTableJobInput | TritonModelJobInput | UriFileJobInput | UriFolderJobInput}
- Inputs for the pipeline job.
- isArchived
boolean
- Is the asset archived?
- jobs
{[key: string]: any}
- Jobs construct the Pipeline Job.
- notificationSetting
NotificationSetting
- Notification setting for the job
- outputs
{[key: string]: CustomModelJobOutput | MLFlowModelJobOutput | MLTableJobOutput | TritonModelJobOutput | UriFileJobOutput | UriFolderJobOutput}
- Outputs for the pipeline job
- properties
{[key: string]: string}
- The asset property dictionary.
- services
{[key: string]: JobService}
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- settings
any
- Pipeline settings, for things like ContinueRunOnStepFailure etc.
- sourceJobId
string
- ARM resource ID of source job.
- tags
{[key: string]: string}
- Tag dictionary. Tags can be added, removed, and updated.
- component_id
str
- ARM resource ID of the component resource.
- compute_id
str
- ARM resource ID of the compute resource.
- description
str
- The asset description text.
- display_name
str
- Display name of job.
- experiment_name
str
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlToken | ManagedIdentity | UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Mapping[str, Union[CustomModelJobInput, LiteralJobInput, MLFlowModelJobInput, MLTableJobInput, TritonModelJobInput, UriFileJobInput, UriFolderJobInput]]
- Inputs for the pipeline job.
- is_archived
bool
- Is the asset archived?
- jobs
Mapping[str, Any]
- Jobs construct the Pipeline Job.
- notification_setting
NotificationSetting
- Notification setting for the job
- outputs
Mapping[str, Union[CustomModelJobOutput, MLFlowModelJobOutput, MLTableJobOutput, TritonModelJobOutput, UriFileJobOutput, UriFolderJobOutput]]
- Outputs for the pipeline job
- properties
Mapping[str, str]
- The asset property dictionary.
- services
Mapping[str, JobService]
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- settings
Any
- Pipeline settings, for things like ContinueRunOnStepFailure etc.
- source_job_id
str
- ARM resource ID of source job.
- tags
Mapping[str, str]
- Tag dictionary. Tags can be added, removed, and updated.
- 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
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
Property Map | Property Map | Property Map
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Map<Property Map | Property Map | Property Map | Property Map | Property Map | Property Map | Property Map>
- Inputs for the pipeline job.
- isArchived
Boolean
- Is the asset archived?
- jobs
Map<Any>
- Jobs construct the Pipeline Job.
- notificationSetting
Property Map
- Notification setting for the job
- outputs
Map<Property Map | Property Map | Property Map | Property Map | Property Map | Property Map>
- Outputs for the pipeline job
- properties
Map<String>
- The asset property dictionary.
- services
Map<Property Map>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- settings
Any
- Pipeline settings, for things like ContinueRunOnStepFailure etc.
- sourceJobId
String
- ARM resource ID of source job.
- tags
Map<String>
- Tag dictionary. Tags can be added, removed, and updated.
PipelineJobResponse
Pipeline Job definition: defines generic to MFE attributes.
- Status
string
- Status of the job.
- 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
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- Identity
Pulumi.AzureNative.MachineLearningServices.Inputs.AmlTokenResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.ManagedIdentityResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- Inputs
Dictionary<string, object>
- Inputs for the pipeline job.
- IsArchived
bool
- Is the asset archived?
- Jobs
Dictionary<string, object>
- Jobs construct the Pipeline Job.
- NotificationSetting
Pulumi.AzureNative.MachineLearningServices.Inputs.NotificationSettingResponse
- Notification setting for the job
- Outputs
Dictionary<string, object>
- Outputs for the pipeline job
- Properties
Dictionary<string, string>
- The asset property dictionary.
- Services
Dictionary<string, Pulumi.AzureNative.MachineLearningServices.Inputs.JobServiceResponse>
- 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.
- Tags
Dictionary<string, string>
- Tag dictionary. Tags can be added, removed, and updated.
- Status
string
- Status of the job.
- 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
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- Identity
AmlTokenResponse | ManagedIdentityResponse | UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- Inputs
map[string]interface{}
- Inputs for the pipeline job.
- IsArchived
bool
- Is the asset archived?
- Jobs
map[string]interface{}
- Jobs construct the Pipeline Job.
- NotificationSetting
NotificationSettingResponse
- Notification setting for the job
- Outputs
map[string]interface{}
- Outputs for the pipeline job
- Properties
map[string]string
- The asset property dictionary.
- Services
map[string]JobServiceResponse
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- Settings
interface{}
- Pipeline settings, for things like ContinueRunOnStepFailure etc.
- SourceJobId
string
- ARM resource ID of source job.
- Tags
map[string]string
- Tag dictionary. Tags can be added, removed, and updated.
- status
string
- Status of the job.
- component_id
string
- ARM resource ID of the component resource.
- compute_id
string
- ARM resource ID of the compute resource.
- description
string
- The asset description text.
- display_name
string
- Display name of job.
- experiment_name
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
object | object | object
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
map(object | object | object | object | object | object | object)
- Inputs for the pipeline job.
- is_archived
bool
- Is the asset archived?
- jobs
map(any)
- Jobs construct the Pipeline Job.
- notification_setting
object
- Notification setting for the job
- outputs
map(object | object | object | object | object | object)
- Outputs for the pipeline job
- properties
map(string)
- The asset property dictionary.
- services
map(object)
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- settings
any
- Pipeline settings, for things like ContinueRunOnStepFailure etc.
- source_job_id
string
- ARM resource ID of source job.
- tags
map(string)
- Tag dictionary. Tags can be added, removed, and updated.
- status
String
- Status of the job.
- 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
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlTokenResponse | ManagedIdentityResponse | UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Map<String,Object>
- Inputs for the pipeline job.
- isArchived
Boolean
- Is the asset archived?
- jobs
Map<String,Object>
- Jobs construct the Pipeline Job.
- notificationSetting
NotificationSettingResponse
- Notification setting for the job
- outputs
Map<String,Object>
- Outputs for the pipeline job
- properties
Map<String,String>
- The asset property dictionary.
- services
Map<String,JobServiceResponse>
- 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.
- tags
Map<String,String>
- Tag dictionary. Tags can be added, removed, and updated.
- status
string
- Status of the job.
- 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
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlTokenResponse | ManagedIdentityResponse | UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
{[key: string]: CustomModelJobInputResponse | LiteralJobInputResponse | MLFlowModelJobInputResponse | MLTableJobInputResponse | TritonModelJobInputResponse | UriFileJobInputResponse | UriFolderJobInputResponse}
- Inputs for the pipeline job.
- isArchived
boolean
- Is the asset archived?
- jobs
{[key: string]: any}
- Jobs construct the Pipeline Job.
- notificationSetting
NotificationSettingResponse
- Notification setting for the job
- outputs
{[key: string]: CustomModelJobOutputResponse | MLFlowModelJobOutputResponse | MLTableJobOutputResponse | TritonModelJobOutputResponse | UriFileJobOutputResponse | UriFolderJobOutputResponse}
- Outputs for the pipeline job
- properties
{[key: string]: string}
- The asset property dictionary.
- services
{[key: string]: JobServiceResponse}
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- settings
any
- Pipeline settings, for things like ContinueRunOnStepFailure etc.
- sourceJobId
string
- ARM resource ID of source job.
- tags
{[key: string]: string}
- Tag dictionary. Tags can be added, removed, and updated.
- status
str
- Status of the job.
- component_id
str
- ARM resource ID of the component resource.
- compute_id
str
- ARM resource ID of the compute resource.
- description
str
- The asset description text.
- display_name
str
- Display name of job.
- experiment_name
str
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlTokenResponse | ManagedIdentityResponse | UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Mapping[str, Union[CustomModelJobInputResponse, LiteralJobInputResponse, MLFlowModelJobInputResponse, MLTableJobInputResponse, TritonModelJobInputResponse, UriFileJobInputResponse, UriFolderJobInputResponse]]
- Inputs for the pipeline job.
- is_archived
bool
- Is the asset archived?
- jobs
Mapping[str, Any]
- Jobs construct the Pipeline Job.
- notification_setting
NotificationSettingResponse
- Notification setting for the job
- outputs
Mapping[str, Union[CustomModelJobOutputResponse, MLFlowModelJobOutputResponse, MLTableJobOutputResponse, TritonModelJobOutputResponse, UriFileJobOutputResponse, UriFolderJobOutputResponse]]
- Outputs for the pipeline job
- properties
Mapping[str, str]
- The asset property dictionary.
- services
Mapping[str, JobServiceResponse]
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- settings
Any
- Pipeline settings, for things like ContinueRunOnStepFailure etc.
- source_job_id
str
- ARM resource ID of source job.
- tags
Mapping[str, str]
- Tag dictionary. Tags can be added, removed, and updated.
- status
String
- Status of the job.
- 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
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
Property Map | Property Map | Property Map
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Map<Property Map | Property Map | Property Map | Property Map | Property Map | Property Map | Property Map>
- Inputs for the pipeline job.
- isArchived
Boolean
- Is the asset archived?
- jobs
Map<Any>
- Jobs construct the Pipeline Job.
- notificationSetting
Property Map
- Notification setting for the job
- outputs
Map<Property Map | Property Map | Property Map | Property Map | Property Map | Property Map>
- Outputs for the pipeline job
- properties
Map<String>
- The asset property dictionary.
- services
Map<Property Map>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- settings
Any
- Pipeline settings, for things like ContinueRunOnStepFailure etc.
- sourceJobId
String
- ARM resource ID of source job.
- tags
Map<String>
- Tag dictionary. Tags can be added, removed, and updated.
PredictionDriftMonitoringSignal, PredictionDriftMonitoringSignalArgs
PredictionDriftMonitoringSignalResponse
PyTorch, PyTorchArgs
PyTorch distribution configuration.
PyTorchResponse
PyTorch distribution configuration.
QueueSettings, QueueSettingsArgs
QueueSettingsResponse
- JobTier
string
- Enum to determine the job tier.
- JobTier
string
- Enum to determine the job tier.
- job_tier
string
- Enum to determine the job tier.
- jobTier
String
- Enum to determine the job tier.
- jobTier
string
- Enum to determine the job tier.
- job_tier
str
- Enum to determine the job tier.
- jobTier
String
- Enum to determine the job tier.
RandomSamplingAlgorithm, RandomSamplingAlgorithmArgs
Defines a Sampling Algorithm that generates values randomly
- rule
string | "Random" | "Sobol"
- The specific type of random algorithm
- seed
number
- An optional integer to use as the seed for random number generation
- rule
String | "Random" | "Sobol"
- The specific type of random algorithm
- seed
Number
- An optional integer to use as the seed for random number generation
RandomSamplingAlgorithmResponse
Defines a Sampling Algorithm that generates values randomly
- Rule
string
- The specific type of random algorithm
- Seed
int
- An optional integer to use as the seed for random number generation
- Rule
string
- The specific type of random algorithm
- Seed
int
- An optional integer to use as the seed for random number generation
- rule
string
- The specific type of random algorithm
- seed
number
- An optional integer to use as the seed for random number generation
- rule
String
- The specific type of random algorithm
- seed
Integer
- An optional integer to use as the seed for random number generation
- rule
string
- The specific type of random algorithm
- seed
number
- An optional integer to use as the seed for random number generation
- rule
str
- The specific type of random algorithm
- seed
int
- An optional integer to use as the seed for random number generation
- rule
String
- The specific type of random algorithm
- seed
Number
- An optional integer to use as the seed for random number generation
RandomSamplingAlgorithmRule
The specific type of random algorithm
| Name | Value |
|---|
| Random | Random |
| Sobol | Sobol |
| Name | Value |
|---|
| RandomSamplingAlgorithmRuleRandom | Random |
| RandomSamplingAlgorithmRuleSobol | Sobol |
| Name | Value |
|---|
| Random | Random |
| Sobol | Sobol |
| Name | Value |
|---|
| Random | Random |
| Sobol | Sobol |
| Name | Value |
|---|
| RANDOM | Random |
| SOBOL | Sobol |
RecurrenceFrequency
[Required] The frequency to trigger schedule.
| Name | Value | Description |
|---|
| Minute | Minute | Minute frequency |
| Hour | Hour | Hour frequency |
| Day | Day | Day frequency |
| Week | Week | Week frequency |
| Month | Month | Month frequency |
| Name | Value | Description |
|---|
| RecurrenceFrequencyMinute | Minute | Minute frequency |
| RecurrenceFrequencyHour | Hour | Hour frequency |
| RecurrenceFrequencyDay | Day | Day frequency |
| RecurrenceFrequencyWeek | Week | Week frequency |
| RecurrenceFrequencyMonth | Month | Month frequency |
| Value | Description |
|---|
Minute | Minute frequency |
Hour | Hour frequency |
Day | Day frequency |
Week | Week frequency |
Month | Month frequency |
| Name | Value | Description |
|---|
| Minute | Minute | Minute frequency |
| Hour | Hour | Hour frequency |
| Day | Day | Day frequency |
| Week | Week | Week frequency |
| Month | Month | Month frequency |
| Name | Value | Description |
|---|
| Minute | Minute | Minute frequency |
| Hour | Hour | Hour frequency |
| Day | Day | Day frequency |
| Week | Week | Week frequency |
| Month | Month | Month frequency |
| Name | Value | Description |
|---|
| MINUTE | Minute | Minute frequency |
| HOUR | Hour | Hour frequency |
| DAY | Day | Day frequency |
| WEEK | Week | Week frequency |
| MONTH | Month | Month frequency |
| Value | Description |
|---|
Minute | Minute frequency |
Hour | Hour frequency |
Day | Day frequency |
Week | Week frequency |
Month | Month frequency |
RecurrenceSchedule, RecurrenceScheduleArgs
- Hours
List<int>
- [Required] List of hours for the schedule.
- Minutes
List<int>
- [Required] List of minutes for the schedule.
- MonthDays
List<int>
- List of month days for the schedule
- WeekDays
List<Union<string, Pulumi.AzureNative.MachineLearningServices.WeekDay>>
- List of days for the schedule.
- Hours
[]int
- [Required] List of hours for the schedule.
- Minutes
[]int
- [Required] List of minutes for the schedule.
- MonthDays
[]int
- List of month days for the schedule
- WeekDays
[]string
- List of days for the schedule.
- hours
list(number)
- [Required] List of hours for the schedule.
- minutes
list(number)
- [Required] List of minutes for the schedule.
- month_days
list(number)
- List of month days for the schedule
- week_days
list(string | "Monday" | "Tuesday" | "Wednesday" | "Thursday" | "Friday" | "Saturday" | "Sunday")
- List of days for the schedule.
- hours
List<Integer>
- [Required] List of hours for the schedule.
- minutes
List<Integer>
- [Required] List of minutes for the schedule.
- monthDays
List<Integer>
- List of month days for the schedule
- weekDays
List<Either<String,WeekDay>>
- List of days for the schedule.
- hours
number[]
- [Required] List of hours for the schedule.
- minutes
number[]
- [Required] List of minutes for the schedule.
- monthDays
number[]
- List of month days for the schedule
- weekDays
(string | WeekDay)[]
- List of days for the schedule.
- hours
Sequence[int]
- [Required] List of hours for the schedule.
- minutes
Sequence[int]
- [Required] List of minutes for the schedule.
- month_days
Sequence[int]
- List of month days for the schedule
- week_days
Sequence[Union[str, WeekDay]]
- List of days for the schedule.
- hours
List<Number>
- [Required] List of hours for the schedule.
- minutes
List<Number>
- [Required] List of minutes for the schedule.
- monthDays
List<Number>
- List of month days for the schedule
- weekDays
List<String | "Monday" | "Tuesday" | "Wednesday" | "Thursday" | "Friday" | "Saturday" | "Sunday">
- List of days for the schedule.
RecurrenceScheduleResponse
- Hours
List<int>
- [Required] List of hours for the schedule.
- Minutes
List<int>
- [Required] List of minutes for the schedule.
- MonthDays
List<int>
- List of month days for the schedule
- WeekDays
List<string>
- List of days for the schedule.
- Hours
[]int
- [Required] List of hours for the schedule.
- Minutes
[]int
- [Required] List of minutes for the schedule.
- MonthDays
[]int
- List of month days for the schedule
- WeekDays
[]string
- List of days for the schedule.
- hours
list(number)
- [Required] List of hours for the schedule.
- minutes
list(number)
- [Required] List of minutes for the schedule.
- month_days
list(number)
- List of month days for the schedule
- week_days
list(string)
- List of days for the schedule.
- hours
List<Integer>
- [Required] List of hours for the schedule.
- minutes
List<Integer>
- [Required] List of minutes for the schedule.
- monthDays
List<Integer>
- List of month days for the schedule
- weekDays
List<String>
- List of days for the schedule.
- hours
number[]
- [Required] List of hours for the schedule.
- minutes
number[]
- [Required] List of minutes for the schedule.
- monthDays
number[]
- List of month days for the schedule
- weekDays
string[]
- List of days for the schedule.
- hours
Sequence[int]
- [Required] List of hours for the schedule.
- minutes
Sequence[int]
- [Required] List of minutes for the schedule.
- month_days
Sequence[int]
- List of month days for the schedule
- week_days
Sequence[str]
- List of days for the schedule.
- hours
List<Number>
- [Required] List of hours for the schedule.
- minutes
List<Number>
- [Required] List of minutes for the schedule.
- monthDays
List<Number>
- List of month days for the schedule
- weekDays
List<String>
- List of days for the schedule.
RecurrenceTrigger, RecurrenceTriggerArgs
RecurrenceTriggerResponse
Regression, RegressionArgs
Regression task in AutoML Table vertical.
RegressionModels
Enum for all Regression models supported by AutoML.
| Name | Value | Description |
|---|
| ElasticNet | ElasticNet | Elastic net is a popular type of regularized linear regression that combines two popular penalties, specifically the L1 and L2 penalty functions. |
| GradientBoosting | GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
| DecisionTree | DecisionTree | 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. |
| KNN | KNN | 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 | LassoLars | Lasso model fit with Least Angle Regression a.k.a. Lars. It is a Linear Model trained with an L1 prior as regularizer. |
| SGD | SGD | 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. |
| RandomForest | RandomForest | 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. |
| ExtremeRandomTrees | ExtremeRandomTrees | 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. |
| LightGBM | LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
| XGBoostRegressor | XGBoostRegressor | XGBoostRegressor: Extreme Gradient Boosting Regressor is a supervised machine learning model using ensemble of base learners. |
| Name | Value | Description |
|---|
| RegressionModelsElasticNet | ElasticNet | Elastic net is a popular type of regularized linear regression that combines two popular penalties, specifically the L1 and L2 penalty functions. |
| RegressionModelsGradientBoosting | GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
| RegressionModelsDecisionTree | DecisionTree | 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. |
| RegressionModelsKNN | KNN | 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. |
| RegressionModelsLassoLars | LassoLars | Lasso model fit with Least Angle Regression a.k.a. Lars. It is a Linear Model trained with an L1 prior as regularizer. |
| RegressionModelsSGD | SGD | 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. |
| RegressionModelsRandomForest | RandomForest | 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. |
| RegressionModelsExtremeRandomTrees | ExtremeRandomTrees | 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. |
| RegressionModelsLightGBM | LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
| RegressionModelsXGBoostRegressor | XGBoostRegressor | XGBoostRegressor: Extreme Gradient Boosting Regressor is a supervised machine learning model using ensemble of base learners. |
| Value | Description |
|---|
ElasticNet | Elastic net is a popular type of regularized linear regression that combines two popular penalties, specifically the L1 and L2 penalty functions. |
GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
DecisionTree | 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. |
KNN | 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 | Lasso model fit with Least Angle Regression a.k.a. Lars. It is a Linear Model trained with an L1 prior as regularizer. |
SGD | 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. |
RandomForest | 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. |
ExtremeRandomTrees | 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. |
LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
XGBoostRegressor | XGBoostRegressor: Extreme Gradient Boosting Regressor is a supervised machine learning model using ensemble of base learners. |
| Name | Value | Description |
|---|
| ElasticNet | ElasticNet | Elastic net is a popular type of regularized linear regression that combines two popular penalties, specifically the L1 and L2 penalty functions. |
| GradientBoosting | GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
| DecisionTree | DecisionTree | 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. |
| KNN | KNN | 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 | LassoLars | Lasso model fit with Least Angle Regression a.k.a. Lars. It is a Linear Model trained with an L1 prior as regularizer. |
| SGD | SGD | 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. |
| RandomForest | RandomForest | 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. |
| ExtremeRandomTrees | ExtremeRandomTrees | 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. |
| LightGBM | LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
| XGBoostRegressor | XGBoostRegressor | XGBoostRegressor: Extreme Gradient Boosting Regressor is a supervised machine learning model using ensemble of base learners. |
| Name | Value | Description |
|---|
| ElasticNet | ElasticNet | Elastic net is a popular type of regularized linear regression that combines two popular penalties, specifically the L1 and L2 penalty functions. |
| GradientBoosting | GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
| DecisionTree | DecisionTree | 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. |
| KNN | KNN | 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 | LassoLars | Lasso model fit with Least Angle Regression a.k.a. Lars. It is a Linear Model trained with an L1 prior as regularizer. |
| SGD | SGD | 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. |
| RandomForest | RandomForest | 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. |
| ExtremeRandomTrees | ExtremeRandomTrees | 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. |
| LightGBM | LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
| XGBoostRegressor | XGBoostRegressor | XGBoostRegressor: Extreme Gradient Boosting Regressor is a supervised machine learning model using ensemble of base learners. |
| Name | Value | Description |
|---|
| ELASTIC_NET | ElasticNet | Elastic net is a popular type of regularized linear regression that combines two popular penalties, specifically the L1 and L2 penalty functions. |
| GRADIENT_BOOSTING | GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
| DECISION_TREE | DecisionTree | 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. |
| KNN | KNN | 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. |
| LASSO_LARS | LassoLars | Lasso model fit with Least Angle Regression a.k.a. Lars. It is a Linear Model trained with an L1 prior as regularizer. |
| SGD | SGD | 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. |
| RANDOM_FOREST | RandomForest | 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. |
| EXTREME_RANDOM_TREES | ExtremeRandomTrees | 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. |
| LIGHT_GBM | LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
| XG_BOOST_REGRESSOR | XGBoostRegressor | XGBoostRegressor: Extreme Gradient Boosting Regressor is a supervised machine learning model using ensemble of base learners. |
| Value | Description |
|---|
ElasticNet | Elastic net is a popular type of regularized linear regression that combines two popular penalties, specifically the L1 and L2 penalty functions. |
GradientBoosting | The technique of transiting week learners into a strong learner is called Boosting. The gradient boosting algorithm process works on this theory of execution. |
DecisionTree | 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. |
KNN | 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 | Lasso model fit with Least Angle Regression a.k.a. Lars. It is a Linear Model trained with an L1 prior as regularizer. |
SGD | 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. |
RandomForest | 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. |
ExtremeRandomTrees | 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. |
LightGBM | LightGBM is a gradient boosting framework that uses tree based learning algorithms. |
XGBoostRegressor | XGBoostRegressor: Extreme Gradient Boosting Regressor is a supervised machine learning model using ensemble of base learners. |
RegressionPrimaryMetrics
Primary metrics for Regression task.
| Name | Value | Description |
|---|
| SpearmanCorrelation | SpearmanCorrelation | The Spearman's rank coefficient of correlation is a nonparametric measure of rank correlation. |
| NormalizedRootMeanSquaredError | NormalizedRootMeanSquaredError | The Normalized Root Mean Squared Error (NRMSE) the RMSE facilitates the comparison between models with different scales. |
| R2Score | R2Score | The R2 score is one of the performance evaluation measures for forecasting-based machine learning models. |
| NormalizedMeanAbsoluteError | NormalizedMeanAbsoluteError | The Normalized Mean Absolute Error (NMAE) is a validation metric to compare the Mean Absolute Error (MAE) of (time) series with different scales. |
| Name | Value | Description |
|---|
| RegressionPrimaryMetricsSpearmanCorrelation | SpearmanCorrelation | The Spearman's rank coefficient of correlation is a nonparametric measure of rank correlation. |
| RegressionPrimaryMetricsNormalizedRootMeanSquaredError | NormalizedRootMeanSquaredError | The Normalized Root Mean Squared Error (NRMSE) the RMSE facilitates the comparison between models with different scales. |
| RegressionPrimaryMetricsR2Score | R2Score | The R2 score is one of the performance evaluation measures for forecasting-based machine learning models. |
| RegressionPrimaryMetricsNormalizedMeanAbsoluteError | NormalizedMeanAbsoluteError | The Normalized Mean Absolute Error (NMAE) is a validation metric to compare the Mean Absolute Error (MAE) of (time) series with different scales. |
| Value | Description |
|---|
SpearmanCorrelation | The Spearman's rank coefficient of correlation is a nonparametric measure of rank correlation. |
NormalizedRootMeanSquaredError | The Normalized Root Mean Squared Error (NRMSE) the RMSE facilitates the comparison between models with different scales. |
R2Score | The R2 score is one of the performance evaluation measures for forecasting-based machine learning models. |
NormalizedMeanAbsoluteError | The Normalized Mean Absolute Error (NMAE) is a validation metric to compare the Mean Absolute Error (MAE) of (time) series with different scales. |
| Name | Value | Description |
|---|
| SpearmanCorrelation | SpearmanCorrelation | The Spearman's rank coefficient of correlation is a nonparametric measure of rank correlation. |
| NormalizedRootMeanSquaredError | NormalizedRootMeanSquaredError | The Normalized Root Mean Squared Error (NRMSE) the RMSE facilitates the comparison between models with different scales. |
| R2Score | R2Score | The R2 score is one of the performance evaluation measures for forecasting-based machine learning models. |
| NormalizedMeanAbsoluteError | NormalizedMeanAbsoluteError | The Normalized Mean Absolute Error (NMAE) is a validation metric to compare the Mean Absolute Error (MAE) of (time) series with different scales. |
| Name | Value | Description |
|---|
| SpearmanCorrelation | SpearmanCorrelation | The Spearman's rank coefficient of correlation is a nonparametric measure of rank correlation. |
| NormalizedRootMeanSquaredError | NormalizedRootMeanSquaredError | The Normalized Root Mean Squared Error (NRMSE) the RMSE facilitates the comparison between models with different scales. |
| R2Score | R2Score | The R2 score is one of the performance evaluation measures for forecasting-based machine learning models. |
| NormalizedMeanAbsoluteError | NormalizedMeanAbsoluteError | The Normalized Mean Absolute Error (NMAE) is a validation metric to compare the Mean Absolute Error (MAE) of (time) series with different scales. |
| Name | Value | Description |
|---|
| SPEARMAN_CORRELATION | SpearmanCorrelation | The Spearman's rank coefficient of correlation is a nonparametric measure of rank correlation. |
| NORMALIZED_ROOT_MEAN_SQUARED_ERROR | NormalizedRootMeanSquaredError | The Normalized Root Mean Squared Error (NRMSE) the RMSE facilitates the comparison between models with different scales. |
| R2_SCORE | R2Score | The R2 score is one of the performance evaluation measures for forecasting-based machine learning models. |
| NORMALIZED_MEAN_ABSOLUTE_ERROR | NormalizedMeanAbsoluteError | The Normalized Mean Absolute Error (NMAE) is a validation metric to compare the Mean Absolute Error (MAE) of (time) series with different scales. |
| Value | Description |
|---|
SpearmanCorrelation | The Spearman's rank coefficient of correlation is a nonparametric measure of rank correlation. |
NormalizedRootMeanSquaredError | The Normalized Root Mean Squared Error (NRMSE) the RMSE facilitates the comparison between models with different scales. |
R2Score | The R2 score is one of the performance evaluation measures for forecasting-based machine learning models. |
NormalizedMeanAbsoluteError | The Normalized Mean Absolute Error (NMAE) is a validation metric to compare the Mean Absolute Error (MAE) of (time) series with different scales. |
RegressionResponse
Regression task in AutoML Table vertical.
RegressionTrainingSettings, RegressionTrainingSettingsArgs
Regression Training related configuration.
- allowedTrainingAlgorithms
List<String | "ElasticNet" | "GradientBoosting" | "DecisionTree" | "KNN" | "LassoLars" | "SGD" | "RandomForest" | "ExtremeRandomTrees" | "LightGBM" | "XGBoostRegressor">
- Allowed models for regression task.
- blockedTrainingAlgorithms
List<String | "ElasticNet" | "GradientBoosting" | "DecisionTree" | "KNN" | "LassoLars" | "SGD" | "RandomForest" | "ExtremeRandomTrees" | "LightGBM" | "XGBoostRegressor">
- Blocked models for regression task.
- enableDnnTraining
Boolean
- Enable recommendation of DNN models.
- enableModelExplainability
Boolean
- Flag to turn on explainability on best model.
- enableOnnxCompatibleModels
Boolean
- Flag for enabling onnx compatible models.
- enableStackEnsemble
Boolean
- Enable stack ensemble run.
- enableVoteEnsemble
Boolean
- Enable voting ensemble run.
- ensembleModelDownloadTimeout
String
- 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
Property Map
- Stack ensemble settings for stack ensemble run.
RegressionTrainingSettingsResponse
Regression Training related configuration.
Rolling input data definition.
- JobInputType
string | JobInputType
- [Required] Specifies the type of job.
- 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.
- Columns
map[string]string
- Mapping of column names to special uses.
- DataContext
string
- The context metadata of the data source.
- PreprocessingComponentId
string
- Reference to the component asset used to preprocess the data.
- jobInputType
String | JobInputType
- [Required] Specifies the type of job.
- 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.
- columns
Map<String,String>
- Mapping of column names to special uses.
- dataContext
String
- The context metadata of the data source.
- preprocessingComponentId
String
- Reference to the component asset used to preprocess the data.
- jobInputType
string | JobInputType
- [Required] Specifies the type of job.
- 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.
- columns
{[key: string]: string}
- Mapping of column names to special uses.
- dataContext
string
- The context metadata of the data source.
- preprocessingComponentId
string
- Reference to the component asset used to preprocess the data.
- job_input_type
str | JobInputType
- [Required] Specifies the type of job.
- uri
str
- [Required] Input Asset URI.
- window_offset
str
- [Required] The time offset between the end of the data window and the monitor's current run time.
- window_size
str
- [Required] The size of the rolling data window.
- columns
Mapping[str, str]
- Mapping of column names to special uses.
- data_context
str
- The context metadata of the data source.
- preprocessing_component_id
str
- Reference to the component asset used to preprocess the data.
Rolling input data definition.
- JobInputType
string
- [Required] Specifies the type of job.
- 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.
- Columns
Dictionary<string, string>
- Mapping of column names to special uses.
- DataContext
string
- The context metadata of the data source.
- PreprocessingComponentId
string
- Reference to the component asset used to preprocess the data.
- JobInputType
string
- [Required] Specifies the type of job.
- 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.
- Columns
map[string]string
- Mapping of column names to special uses.
- DataContext
string
- The context metadata of the data source.
- PreprocessingComponentId
string
- Reference to the component asset used to preprocess the data.
- job_input_type
string
- [Required] Specifies the type of job.
- uri
string
- [Required] Input Asset URI.
- window_offset
string
- [Required] The time offset between the end of the data window and the monitor's current run time.
- window_size
string
- [Required] The size of the rolling data window.
- columns
map(string)
- Mapping of column names to special uses.
- data_context
string
- The context metadata of the data source.
- preprocessing_component_id
string
- Reference to the component asset used to preprocess the data.
- jobInputType
String
- [Required] Specifies the type of job.
- 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.
- columns
Map<String,String>
- Mapping of column names to special uses.
- dataContext
String
- The context metadata of the data source.
- preprocessingComponentId
String
- Reference to the component asset used to preprocess the data.
- jobInputType
string
- [Required] Specifies the type of job.
- 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.
- columns
{[key: string]: string}
- Mapping of column names to special uses.
- dataContext
string
- The context metadata of the data source.
- preprocessingComponentId
string
- Reference to the component asset used to preprocess the data.
- job_input_type
str
- [Required] Specifies the type of job.
- uri
str
- [Required] Input Asset URI.
- window_offset
str
- [Required] The time offset between the end of the data window and the monitor's current run time.
- window_size
str
- [Required] The size of the rolling data window.
- columns
Mapping[str, str]
- Mapping of column names to special uses.
- data_context
str
- The context metadata of the data source.
- preprocessing_component_id
str
- Reference to the component asset used to preprocess the data.
- jobInputType
String
- [Required] Specifies the type of job.
- 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.
- columns
Map<String>
- Mapping of column names to special uses.
- dataContext
String
- The context metadata of the data source.
- preprocessingComponentId
String
- Reference to the component asset used to preprocess the data.
SamplingAlgorithmType
[Required] The algorithm used for generating hyperparameter values, along with configuration properties
| Name | Value |
|---|
| Grid | Grid |
| Random | Random |
| Bayesian | Bayesian |
| Name | Value |
|---|
| SamplingAlgorithmTypeGrid | Grid |
| SamplingAlgorithmTypeRandom | Random |
| SamplingAlgorithmTypeBayesian | Bayesian |
| Name | Value |
|---|
| Grid | Grid |
| Random | Random |
| Bayesian | Bayesian |
| Name | Value |
|---|
| Grid | Grid |
| Random | Random |
| Bayesian | Bayesian |
| Name | Value |
|---|
| GRID | Grid |
| RANDOM | Random |
| BAYESIAN | Bayesian |
ScheduleProperties, SchedulePropertiesArgs
Base definition of a schedule
SchedulePropertiesResponse
Base definition of a schedule
ShortSeriesHandlingConfiguration
The parameter defining how if AutoML should handle short time series.
| Name | Value | Description |
|---|
| None | None | Represents no/null value. |
| Auto | Auto | Short series will be padded if there are no long series, otherwise short series will be dropped. |
| Pad | Pad | All the short series will be padded. |
| Drop | Drop | All the short series will be dropped. |
| Name | Value | Description |
|---|
| ShortSeriesHandlingConfigurationNone | None | Represents no/null value. |
| ShortSeriesHandlingConfigurationAuto | Auto | Short series will be padded if there are no long series, otherwise short series will be dropped. |
| ShortSeriesHandlingConfigurationPad | Pad | All the short series will be padded. |
| ShortSeriesHandlingConfigurationDrop | Drop | All the short series will be dropped. |
| Value | Description |
|---|
None | Represents no/null value. |
Auto | Short series will be padded if there are no long series, otherwise short series will be dropped. |
Pad | All the short series will be padded. |
Drop | All the short series will be dropped. |
| Name | Value | Description |
|---|
| None | None | Represents no/null value. |
| Auto | Auto | Short series will be padded if there are no long series, otherwise short series will be dropped. |
| Pad | Pad | All the short series will be padded. |
| Drop | Drop | All the short series will be dropped. |
| Name | Value | Description |
|---|
| None | None | Represents no/null value. |
| Auto | Auto | Short series will be padded if there are no long series, otherwise short series will be dropped. |
| Pad | Pad | All the short series will be padded. |
| Drop | Drop | All the short series will be dropped. |
| Name | Value | Description |
|---|
| NONE | None | Represents no/null value. |
| AUTO | Auto | Short series will be padded if there are no long series, otherwise short series will be dropped. |
| PAD | Pad | All the short series will be padded. |
| DROP | Drop | All the short series will be dropped. |
| Value | Description |
|---|
None | Represents no/null value. |
Auto | Short series will be padded if there are no long series, otherwise short series will be dropped. |
Pad | All the short series will be padded. |
Drop | All the short series will be dropped. |
SparkJob, SparkJobArgs
Spark job definition.
- CodeId
string
- [Required] arm-id of the code asset.
- Entry
Pulumi.AzureNative.MachineLearningServices.Inputs.SparkJobPythonEntry | Pulumi.AzureNative.MachineLearningServices.Inputs.SparkJobScalaEntry
- [Required] The entry to execute on startup of the job.
- Archives
List<string>
- Archive files used in the job.
- Args
string
- Arguments for the job.
- ComponentId
string
- ARM resource ID of the component resource.
- ComputeId
string
- ARM resource ID of the compute resource.
- Conf
Dictionary<string, string>
- Spark configured properties.
- 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.
- EnvironmentVariables
Dictionary<string, string>
- Environment variables included in the job.
- ExperimentName
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- Files
List<string>
- Files used in the job.
- Identity
Pulumi.AzureNative.MachineLearningServices.Inputs.AmlToken | Pulumi.AzureNative.MachineLearningServices.Inputs.ManagedIdentity | Pulumi.AzureNative.MachineLearningServices.Inputs.UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- Inputs
Dictionary<string, object>
- Mapping of input data bindings used in the job.
- IsArchived
bool
- Is the asset archived?
- Jars
List<string>
- Jar files used in the job.
- NotificationSetting
Pulumi.AzureNative.MachineLearningServices.Inputs.NotificationSetting
- Notification setting for the job
- Outputs
Dictionary<string, object>
- Mapping of output data bindings used in the job.
- Properties
Dictionary<string, string>
- The asset property dictionary.
- PyFiles
List<string>
- Python files used in the job.
- QueueSettings
Pulumi.AzureNative.MachineLearningServices.Inputs.QueueSettings
- Queue settings for the job
- Resources
Pulumi.AzureNative.MachineLearningServices.Inputs.SparkResourceConfiguration
- Compute Resource configuration for the job.
- Services
Dictionary<string, Pulumi.AzureNative.MachineLearningServices.Inputs.JobService>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- Tags
Dictionary<string, string>
- Tag dictionary. Tags can be added, removed, and updated.
- code_id
string
- [Required] arm-id of the code asset.
- entry
object | object
- [Required] The entry to execute on startup of the job.
- archives
list(string)
- Archive files used in the job.
- args
string
- Arguments for the job.
- component_id
string
- ARM resource ID of the component resource.
- compute_id
string
- ARM resource ID of the compute resource.
- conf
map(string)
- Spark configured properties.
- description
string
- The asset description text.
- display_name
string
- Display name of job.
- environment_id
string
- The ARM resource ID of the Environment specification for the job.
- environment_variables
map(string)
- Environment variables included in the job.
- experiment_name
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- files
list(string)
- Files used in the job.
- identity
object | object | object
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
map(object | object | object | object | object | object | object)
- Mapping of input data bindings used in the job.
- is_archived
bool
- Is the asset archived?
- jars
list(string)
- Jar files used in the job.
- notification_setting
object
- Notification setting for the job
- outputs
map(object | object | object | object | object | object)
- Mapping of output data bindings used in the job.
- properties
map(string)
- The asset property dictionary.
- py_files
list(string)
- Python files used in the job.
- queue_settings
object
- Queue settings for the job
- resources
object
- Compute Resource configuration for the job.
- services
map(object)
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
map(string)
- Tag dictionary. Tags can be added, removed, and updated.
- codeId
String
- [Required] arm-id of the code asset.
- entry
SparkJobPythonEntry | SparkJobScalaEntry
- [Required] The entry to execute on startup of the job.
- archives
List<String>
- Archive files used in the job.
- args
String
- Arguments for the job.
- componentId
String
- ARM resource ID of the component resource.
- computeId
String
- ARM resource ID of the compute resource.
- conf
Map<String,String>
- Spark configured properties.
- 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.
- environmentVariables
Map<String,String>
- Environment variables included in the job.
- experimentName
String
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- files
List<String>
- Files used in the job.
- identity
AmlToken | ManagedIdentity | UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Map<String,Object>
- Mapping of input data bindings used in the job.
- isArchived
Boolean
- Is the asset archived?
- jars
List<String>
- Jar files used in the job.
- notificationSetting
NotificationSetting
- Notification setting for the job
- outputs
Map<String,Object>
- Mapping of output data bindings used in the job.
- properties
Map<String,String>
- The asset property dictionary.
- pyFiles
List<String>
- Python files used in the job.
- queueSettings
QueueSettings
- Queue settings for the job
- resources
SparkResourceConfiguration
- Compute Resource configuration for the job.
- services
Map<String,JobService>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Map<String,String>
- Tag dictionary. Tags can be added, removed, and updated.
- codeId
string
- [Required] arm-id of the code asset.
- entry
SparkJobPythonEntry | SparkJobScalaEntry
- [Required] The entry to execute on startup of the job.
- archives
string[]
- Archive files used in the job.
- args
string
- Arguments for the job.
- componentId
string
- ARM resource ID of the component resource.
- computeId
string
- ARM resource ID of the compute resource.
- conf
{[key: string]: string}
- Spark configured properties.
- 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.
- environmentVariables
{[key: string]: string}
- Environment variables included in the job.
- experimentName
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- files
string[]
- Files used in the job.
- identity
AmlToken | ManagedIdentity | UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
{[key: string]: CustomModelJobInput | LiteralJobInput | MLFlowModelJobInput | MLTableJobInput | TritonModelJobInput | UriFileJobInput | UriFolderJobInput}
- Mapping of input data bindings used in the job.
- isArchived
boolean
- Is the asset archived?
- jars
string[]
- Jar files used in the job.
- notificationSetting
NotificationSetting
- Notification setting for the job
- outputs
{[key: string]: CustomModelJobOutput | MLFlowModelJobOutput | MLTableJobOutput | TritonModelJobOutput | UriFileJobOutput | UriFolderJobOutput}
- Mapping of output data bindings used in the job.
- properties
{[key: string]: string}
- The asset property dictionary.
- pyFiles
string[]
- Python files used in the job.
- queueSettings
QueueSettings
- Queue settings for the job
- resources
SparkResourceConfiguration
- Compute Resource configuration for the job.
- services
{[key: string]: JobService}
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
{[key: string]: string}
- Tag dictionary. Tags can be added, removed, and updated.
- code_id
str
- [Required] arm-id of the code asset.
- entry
SparkJobPythonEntry | SparkJobScalaEntry
- [Required] The entry to execute on startup of the job.
- archives
Sequence[str]
- Archive files used in the job.
- args
str
- Arguments for the job.
- component_id
str
- ARM resource ID of the component resource.
- compute_id
str
- ARM resource ID of the compute resource.
- conf
Mapping[str, str]
- Spark configured properties.
- description
str
- The asset description text.
- display_name
str
- Display name of job.
- environment_id
str
- The ARM resource ID of the Environment specification for the job.
- environment_variables
Mapping[str, str]
- Environment variables included in the job.
- experiment_name
str
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- files
Sequence[str]
- Files used in the job.
- identity
AmlToken | ManagedIdentity | UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Mapping[str, Union[CustomModelJobInput, LiteralJobInput, MLFlowModelJobInput, MLTableJobInput, TritonModelJobInput, UriFileJobInput, UriFolderJobInput]]
- Mapping of input data bindings used in the job.
- is_archived
bool
- Is the asset archived?
- jars
Sequence[str]
- Jar files used in the job.
- notification_setting
NotificationSetting
- Notification setting for the job
- outputs
Mapping[str, Union[CustomModelJobOutput, MLFlowModelJobOutput, MLTableJobOutput, TritonModelJobOutput, UriFileJobOutput, UriFolderJobOutput]]
- Mapping of output data bindings used in the job.
- properties
Mapping[str, str]
- The asset property dictionary.
- py_files
Sequence[str]
- Python files used in the job.
- queue_settings
QueueSettings
- Queue settings for the job
- resources
SparkResourceConfiguration
- Compute Resource configuration for the job.
- services
Mapping[str, JobService]
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Mapping[str, str]
- Tag dictionary. Tags can be added, removed, and updated.
- codeId
String
- [Required] arm-id of the code asset.
- entry
Property Map | Property Map
- [Required] The entry to execute on startup of the job.
- archives
List<String>
- Archive files used in the job.
- args
String
- Arguments for the job.
- componentId
String
- ARM resource ID of the component resource.
- computeId
String
- ARM resource ID of the compute resource.
- conf
Map<String>
- Spark configured properties.
- 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.
- environmentVariables
Map<String>
- Environment variables included in the job.
- experimentName
String
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- files
List<String>
- Files used in the job.
- identity
Property Map | Property Map | Property Map
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Map<Property Map | Property Map | Property Map | Property Map | Property Map | Property Map | Property Map>
- Mapping of input data bindings used in the job.
- isArchived
Boolean
- Is the asset archived?
- jars
List<String>
- Jar files used in the job.
- notificationSetting
Property Map
- Notification setting for the job
- outputs
Map<Property Map | Property Map | Property Map | Property Map | Property Map | Property Map>
- Mapping of output data bindings used in the job.
- properties
Map<String>
- The asset property dictionary.
- pyFiles
List<String>
- Python files used in the job.
- queueSettings
Property Map
- Queue settings for the job
- resources
Property Map
- Compute Resource configuration for the job.
- services
Map<Property Map>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Map<String>
- Tag dictionary. Tags can be added, removed, and updated.
SparkJobPythonEntry, SparkJobPythonEntryArgs
- File
string
- [Required] Relative python file path for job entry point.
- File
string
- [Required] Relative python file path for job entry point.
- file
string
- [Required] Relative python file path for job entry point.
- file
String
- [Required] Relative python file path for job entry point.
- file
string
- [Required] Relative python file path for job entry point.
- file
str
- [Required] Relative python file path for job entry point.
- file
String
- [Required] Relative python file path for job entry point.
SparkJobPythonEntryResponse
- File
string
- [Required] Relative python file path for job entry point.
- File
string
- [Required] Relative python file path for job entry point.
- file
string
- [Required] Relative python file path for job entry point.
- file
String
- [Required] Relative python file path for job entry point.
- file
string
- [Required] Relative python file path for job entry point.
- file
str
- [Required] Relative python file path for job entry point.
- file
String
- [Required] Relative python file path for job entry point.
SparkJobResponse
Spark job definition.
- CodeId
string
- [Required] arm-id of the code asset.
- Entry
Pulumi.AzureNative.MachineLearningServices.Inputs.SparkJobPythonEntryResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.SparkJobScalaEntryResponse
- [Required] The entry to execute on startup of the job.
- Status
string
- Status of the job.
- Archives
List<string>
- Archive files used in the job.
- Args
string
- Arguments for the job.
- ComponentId
string
- ARM resource ID of the component resource.
- ComputeId
string
- ARM resource ID of the compute resource.
- Conf
Dictionary<string, string>
- Spark configured properties.
- 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.
- EnvironmentVariables
Dictionary<string, string>
- Environment variables included in the job.
- ExperimentName
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- Files
List<string>
- Files used in the job.
- Identity
Pulumi.AzureNative.MachineLearningServices.Inputs.AmlTokenResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.ManagedIdentityResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- Inputs
Dictionary<string, object>
- Mapping of input data bindings used in the job.
- IsArchived
bool
- Is the asset archived?
- Jars
List<string>
- Jar files used in the job.
- NotificationSetting
Pulumi.AzureNative.MachineLearningServices.Inputs.NotificationSettingResponse
- Notification setting for the job
- Outputs
Dictionary<string, object>
- Mapping of output data bindings used in the job.
- Properties
Dictionary<string, string>
- The asset property dictionary.
- PyFiles
List<string>
- Python files used in the job.
- QueueSettings
Pulumi.AzureNative.MachineLearningServices.Inputs.QueueSettingsResponse
- Queue settings for the job
- Resources
Pulumi.AzureNative.MachineLearningServices.Inputs.SparkResourceConfigurationResponse
- Compute Resource configuration for the job.
- Services
Dictionary<string, Pulumi.AzureNative.MachineLearningServices.Inputs.JobServiceResponse>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- Tags
Dictionary<string, string>
- Tag dictionary. Tags can be added, removed, and updated.
- code_id
string
- [Required] arm-id of the code asset.
- entry
object | object
- [Required] The entry to execute on startup of the job.
- status
string
- Status of the job.
- archives
list(string)
- Archive files used in the job.
- args
string
- Arguments for the job.
- component_id
string
- ARM resource ID of the component resource.
- compute_id
string
- ARM resource ID of the compute resource.
- conf
map(string)
- Spark configured properties.
- description
string
- The asset description text.
- display_name
string
- Display name of job.
- environment_id
string
- The ARM resource ID of the Environment specification for the job.
- environment_variables
map(string)
- Environment variables included in the job.
- experiment_name
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- files
list(string)
- Files used in the job.
- identity
object | object | object
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
map(object | object | object | object | object | object | object)
- Mapping of input data bindings used in the job.
- is_archived
bool
- Is the asset archived?
- jars
list(string)
- Jar files used in the job.
- notification_setting
object
- Notification setting for the job
- outputs
map(object | object | object | object | object | object)
- Mapping of output data bindings used in the job.
- properties
map(string)
- The asset property dictionary.
- py_files
list(string)
- Python files used in the job.
- queue_settings
object
- Queue settings for the job
- resources
object
- Compute Resource configuration for the job.
- services
map(object)
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
map(string)
- Tag dictionary. Tags can be added, removed, and updated.
- codeId
String
- [Required] arm-id of the code asset.
- entry
SparkJobPythonEntryResponse | SparkJobScalaEntryResponse
- [Required] The entry to execute on startup of the job.
- status
String
- Status of the job.
- archives
List<String>
- Archive files used in the job.
- args
String
- Arguments for the job.
- componentId
String
- ARM resource ID of the component resource.
- computeId
String
- ARM resource ID of the compute resource.
- conf
Map<String,String>
- Spark configured properties.
- 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.
- environmentVariables
Map<String,String>
- Environment variables included in the job.
- experimentName
String
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- files
List<String>
- Files used in the job.
- identity
AmlTokenResponse | ManagedIdentityResponse | UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Map<String,Object>
- Mapping of input data bindings used in the job.
- isArchived
Boolean
- Is the asset archived?
- jars
List<String>
- Jar files used in the job.
- notificationSetting
NotificationSettingResponse
- Notification setting for the job
- outputs
Map<String,Object>
- Mapping of output data bindings used in the job.
- properties
Map<String,String>
- The asset property dictionary.
- pyFiles
List<String>
- Python files used in the job.
- queueSettings
QueueSettingsResponse
- Queue settings for the job
- resources
SparkResourceConfigurationResponse
- Compute Resource configuration for the job.
- services
Map<String,JobServiceResponse>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Map<String,String>
- Tag dictionary. Tags can be added, removed, and updated.
- codeId
string
- [Required] arm-id of the code asset.
- entry
SparkJobPythonEntryResponse | SparkJobScalaEntryResponse
- [Required] The entry to execute on startup of the job.
- status
string
- Status of the job.
- archives
string[]
- Archive files used in the job.
- args
string
- Arguments for the job.
- componentId
string
- ARM resource ID of the component resource.
- computeId
string
- ARM resource ID of the compute resource.
- conf
{[key: string]: string}
- Spark configured properties.
- 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.
- environmentVariables
{[key: string]: string}
- Environment variables included in the job.
- experimentName
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- files
string[]
- Files used in the job.
- identity
AmlTokenResponse | ManagedIdentityResponse | UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
{[key: string]: CustomModelJobInputResponse | LiteralJobInputResponse | MLFlowModelJobInputResponse | MLTableJobInputResponse | TritonModelJobInputResponse | UriFileJobInputResponse | UriFolderJobInputResponse}
- Mapping of input data bindings used in the job.
- isArchived
boolean
- Is the asset archived?
- jars
string[]
- Jar files used in the job.
- notificationSetting
NotificationSettingResponse
- Notification setting for the job
- outputs
{[key: string]: CustomModelJobOutputResponse | MLFlowModelJobOutputResponse | MLTableJobOutputResponse | TritonModelJobOutputResponse | UriFileJobOutputResponse | UriFolderJobOutputResponse}
- Mapping of output data bindings used in the job.
- properties
{[key: string]: string}
- The asset property dictionary.
- pyFiles
string[]
- Python files used in the job.
- queueSettings
QueueSettingsResponse
- Queue settings for the job
- resources
SparkResourceConfigurationResponse
- Compute Resource configuration for the job.
- services
{[key: string]: JobServiceResponse}
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
{[key: string]: string}
- Tag dictionary. Tags can be added, removed, and updated.
- code_id
str
- [Required] arm-id of the code asset.
- entry
SparkJobPythonEntryResponse | SparkJobScalaEntryResponse
- [Required] The entry to execute on startup of the job.
- status
str
- Status of the job.
- archives
Sequence[str]
- Archive files used in the job.
- args
str
- Arguments for the job.
- component_id
str
- ARM resource ID of the component resource.
- compute_id
str
- ARM resource ID of the compute resource.
- conf
Mapping[str, str]
- Spark configured properties.
- description
str
- The asset description text.
- display_name
str
- Display name of job.
- environment_id
str
- The ARM resource ID of the Environment specification for the job.
- environment_variables
Mapping[str, str]
- Environment variables included in the job.
- experiment_name
str
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- files
Sequence[str]
- Files used in the job.
- identity
AmlTokenResponse | ManagedIdentityResponse | UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Mapping[str, Union[CustomModelJobInputResponse, LiteralJobInputResponse, MLFlowModelJobInputResponse, MLTableJobInputResponse, TritonModelJobInputResponse, UriFileJobInputResponse, UriFolderJobInputResponse]]
- Mapping of input data bindings used in the job.
- is_archived
bool
- Is the asset archived?
- jars
Sequence[str]
- Jar files used in the job.
- notification_setting
NotificationSettingResponse
- Notification setting for the job
- outputs
Mapping[str, Union[CustomModelJobOutputResponse, MLFlowModelJobOutputResponse, MLTableJobOutputResponse, TritonModelJobOutputResponse, UriFileJobOutputResponse, UriFolderJobOutputResponse]]
- Mapping of output data bindings used in the job.
- properties
Mapping[str, str]
- The asset property dictionary.
- py_files
Sequence[str]
- Python files used in the job.
- queue_settings
QueueSettingsResponse
- Queue settings for the job
- resources
SparkResourceConfigurationResponse
- Compute Resource configuration for the job.
- services
Mapping[str, JobServiceResponse]
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Mapping[str, str]
- Tag dictionary. Tags can be added, removed, and updated.
- codeId
String
- [Required] arm-id of the code asset.
- entry
Property Map | Property Map
- [Required] The entry to execute on startup of the job.
- status
String
- Status of the job.
- archives
List<String>
- Archive files used in the job.
- args
String
- Arguments for the job.
- componentId
String
- ARM resource ID of the component resource.
- computeId
String
- ARM resource ID of the compute resource.
- conf
Map<String>
- Spark configured properties.
- 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.
- environmentVariables
Map<String>
- Environment variables included in the job.
- experimentName
String
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- files
List<String>
- Files used in the job.
- identity
Property Map | Property Map | Property Map
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Map<Property Map | Property Map | Property Map | Property Map | Property Map | Property Map | Property Map>
- Mapping of input data bindings used in the job.
- isArchived
Boolean
- Is the asset archived?
- jars
List<String>
- Jar files used in the job.
- notificationSetting
Property Map
- Notification setting for the job
- outputs
Map<Property Map | Property Map | Property Map | Property Map | Property Map | Property Map>
- Mapping of output data bindings used in the job.
- properties
Map<String>
- The asset property dictionary.
- pyFiles
List<String>
- Python files used in the job.
- queueSettings
Property Map
- Queue settings for the job
- resources
Property Map
- Compute Resource configuration for the job.
- services
Map<Property Map>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Map<String>
- Tag dictionary. Tags can be added, removed, and updated.
SparkJobScalaEntry, SparkJobScalaEntryArgs
- ClassName
string
- [Required] Scala class name used as entry point.
- ClassName
string
- [Required] Scala class name used as entry point.
- class_name
string
- [Required] Scala class name used as entry point.
- className
String
- [Required] Scala class name used as entry point.
- className
string
- [Required] Scala class name used as entry point.
- class_name
str
- [Required] Scala class name used as entry point.
- className
String
- [Required] Scala class name used as entry point.
SparkJobScalaEntryResponse
- ClassName
string
- [Required] Scala class name used as entry point.
- ClassName
string
- [Required] Scala class name used as entry point.
- class_name
string
- [Required] Scala class name used as entry point.
- className
String
- [Required] Scala class name used as entry point.
- className
string
- [Required] Scala class name used as entry point.
- class_name
str
- [Required] Scala class name used as entry point.
- className
String
- [Required] Scala class name used as entry point.
SparkResourceConfiguration, SparkResourceConfigurationArgs
- InstanceType
string
- Optional type of VM used as supported by the compute target.
- RuntimeVersion
string
- Version of spark runtime used for the job.
- InstanceType
string
- Optional type of VM used as supported by the compute target.
- RuntimeVersion
string
- Version of spark runtime used for the job.
- instance_type
string
- Optional type of VM used as supported by the compute target.
- runtime_version
string
- Version of spark runtime used for the job.
- instanceType
String
- Optional type of VM used as supported by the compute target.
- runtimeVersion
String
- Version of spark runtime used for the job.
- instanceType
string
- Optional type of VM used as supported by the compute target.
- runtimeVersion
string
- Version of spark runtime used for the job.
- instanceType
String
- Optional type of VM used as supported by the compute target.
- runtimeVersion
String
- Version of spark runtime used for the job.
SparkResourceConfigurationResponse
- InstanceType
string
- Optional type of VM used as supported by the compute target.
- RuntimeVersion
string
- Version of spark runtime used for the job.
- InstanceType
string
- Optional type of VM used as supported by the compute target.
- RuntimeVersion
string
- Version of spark runtime used for the job.
- instance_type
string
- Optional type of VM used as supported by the compute target.
- runtime_version
string
- Version of spark runtime used for the job.
- instanceType
String
- Optional type of VM used as supported by the compute target.
- runtimeVersion
String
- Version of spark runtime used for the job.
- instanceType
string
- Optional type of VM used as supported by the compute target.
- runtimeVersion
string
- Version of spark runtime used for the job.
- instanceType
String
- Optional type of VM used as supported by the compute target.
- runtimeVersion
String
- Version of spark runtime used for the job.
StackEnsembleSettings, StackEnsembleSettingsArgs
Advances setting to customize StackEnsemble run.
- StackMetaLearnerKWargs
object
- Optional parameters to pass to the initializer of the meta-learner.
- StackMetaLearnerTrainPercentage
double
- 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
string | Pulumi.AzureNative.MachineLearningServices.StackMetaLearnerType
- The meta-learner is a model trained on the output of the individual heterogeneous models.\r\nDefault 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).\r\nThis parameter can be one of the following strings: LogisticRegression, LogisticRegressionCV, LightGBMClassifier, ElasticNet, ElasticNetCV, LightGBMRegressor, or LinearRegression
- StackMetaLearnerKWargs
interface{}
- Optional parameters to pass to the initializer of the meta-learner.
- StackMetaLearnerTrainPercentage
float64
- 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
string | StackMetaLearnerType
- The meta-learner is a model trained on the output of the individual heterogeneous models.\r\nDefault 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).\r\nThis parameter can be one of the following strings: LogisticRegression, LogisticRegressionCV, LightGBMClassifier, ElasticNet, ElasticNetCV, LightGBMRegressor, or LinearRegression
- stackMetaLearnerKWargs
Object
- Optional parameters to pass to the initializer of the meta-learner.
- stackMetaLearnerTrainPercentage
Double
- 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
String | StackMetaLearnerType
- The meta-learner is a model trained on the output of the individual heterogeneous models.\r\nDefault 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).\r\nThis parameter can be one of the following strings: LogisticRegression, LogisticRegressionCV, LightGBMClassifier, ElasticNet, ElasticNetCV, LightGBMRegressor, or LinearRegression
- stackMetaLearnerKWargs
any
- Optional parameters to pass to the initializer of the meta-learner.
- stackMetaLearnerTrainPercentage
number
- 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
string | StackMetaLearnerType
- The meta-learner is a model trained on the output of the individual heterogeneous models.\r\nDefault 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).\r\nThis parameter can be one of the following strings: LogisticRegression, LogisticRegressionCV, LightGBMClassifier, ElasticNet, ElasticNetCV, LightGBMRegressor, or LinearRegression
- stack_meta_learner_k_wargs
Any
- Optional parameters to pass to the initializer of the meta-learner.
- stack_meta_learner_train_percentage
float
- 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.
- stack_meta_learner_type
str | StackMetaLearnerType
- The meta-learner is a model trained on the output of the individual heterogeneous models.\r\nDefault 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).\r\nThis parameter can be one of the following strings: LogisticRegression, LogisticRegressionCV, LightGBMClassifier, ElasticNet, ElasticNetCV, LightGBMRegressor, or LinearRegression
StackEnsembleSettingsResponse
Advances setting to customize StackEnsemble run.
- StackMetaLearnerKWargs
object
- Optional parameters to pass to the initializer of the meta-learner.
- StackMetaLearnerTrainPercentage
double
- 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
string
- The meta-learner is a model trained on the output of the individual heterogeneous models.\r\nDefault 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).\r\nThis parameter can be one of the following strings: LogisticRegression, LogisticRegressionCV, LightGBMClassifier, ElasticNet, ElasticNetCV, LightGBMRegressor, or LinearRegression
- StackMetaLearnerKWargs
interface{}
- Optional parameters to pass to the initializer of the meta-learner.
- StackMetaLearnerTrainPercentage
float64
- 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
string
- The meta-learner is a model trained on the output of the individual heterogeneous models.\r\nDefault 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).\r\nThis parameter can be one of the following strings: LogisticRegression, LogisticRegressionCV, LightGBMClassifier, ElasticNet, ElasticNetCV, LightGBMRegressor, or LinearRegression
- stack_meta_learner_k_wargs
any
- Optional parameters to pass to the initializer of the meta-learner.
- stack_meta_learner_train_percentage
number
- 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.
- stack_meta_learner_type
string
- The meta-learner is a model trained on the output of the individual heterogeneous models.\r\nDefault 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).\r\nThis parameter can be one of the following strings: LogisticRegression, LogisticRegressionCV, LightGBMClassifier, ElasticNet, ElasticNetCV, LightGBMRegressor, or LinearRegression
- stackMetaLearnerKWargs
Object
- Optional parameters to pass to the initializer of the meta-learner.
- stackMetaLearnerTrainPercentage
Double
- 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
String
- The meta-learner is a model trained on the output of the individual heterogeneous models.\r\nDefault 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).\r\nThis parameter can be one of the following strings: LogisticRegression, LogisticRegressionCV, LightGBMClassifier, ElasticNet, ElasticNetCV, LightGBMRegressor, or LinearRegression
- stackMetaLearnerKWargs
any
- Optional parameters to pass to the initializer of the meta-learner.
- stackMetaLearnerTrainPercentage
number
- 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
string
- The meta-learner is a model trained on the output of the individual heterogeneous models.\r\nDefault 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).\r\nThis parameter can be one of the following strings: LogisticRegression, LogisticRegressionCV, LightGBMClassifier, ElasticNet, ElasticNetCV, LightGBMRegressor, or LinearRegression
- stack_meta_learner_k_wargs
Any
- Optional parameters to pass to the initializer of the meta-learner.
- stack_meta_learner_train_percentage
float
- 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.
- stack_meta_learner_type
str
- The meta-learner is a model trained on the output of the individual heterogeneous models.\r\nDefault 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).\r\nThis parameter can be one of the following strings: LogisticRegression, LogisticRegressionCV, LightGBMClassifier, ElasticNet, ElasticNetCV, LightGBMRegressor, or LinearRegression
- stackMetaLearnerKWargs
Any
- Optional parameters to pass to the initializer of the meta-learner.
- stackMetaLearnerTrainPercentage
Number
- 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
String
- The meta-learner is a model trained on the output of the individual heterogeneous models.\r\nDefault 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).\r\nThis parameter can be one of the following strings: LogisticRegression, LogisticRegressionCV, LightGBMClassifier, ElasticNet, ElasticNetCV, LightGBMRegressor, or LinearRegression
The meta-learner is a model trained on the output of the individual heterogeneous models.\r\nDefault 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).\r\nThis parameter can be one of the following strings: LogisticRegression, LogisticRegressionCV, LightGBMClassifier, ElasticNet, ElasticNetCV, LightGBMRegressor, or LinearRegression
| Name | Value | Description |
|---|
| None | None | |
| LogisticRegression | LogisticRegression | Default meta-learners are LogisticRegression for classification tasks. |
| LogisticRegressionCV | LogisticRegressionCV | Default meta-learners are LogisticRegression for classification task when CV is on. |
| LightGBMClassifier | LightGBMClassifier | |
| ElasticNet | ElasticNet | Default meta-learners are LogisticRegression for regression task. |
| ElasticNetCV | ElasticNetCV | Default meta-learners are LogisticRegression for regression task when CV is on. |
| LightGBMRegressor | LightGBMRegressor | |
| LinearRegression | LinearRegression | |
| Name | Value | Description |
|---|
| StackMetaLearnerTypeNone | None | |
| StackMetaLearnerTypeLogisticRegression | LogisticRegression | Default meta-learners are LogisticRegression for classification tasks. |
| StackMetaLearnerTypeLogisticRegressionCV | LogisticRegressionCV | Default meta-learners are LogisticRegression for classification task when CV is on. |
| StackMetaLearnerTypeLightGBMClassifier | LightGBMClassifier | |
| StackMetaLearnerTypeElasticNet | ElasticNet | Default meta-learners are LogisticRegression for regression task. |
| StackMetaLearnerTypeElasticNetCV | ElasticNetCV | Default meta-learners are LogisticRegression for regression task when CV is on. |
| StackMetaLearnerTypeLightGBMRegressor | LightGBMRegressor | |
| StackMetaLearnerTypeLinearRegression | LinearRegression | |
| Value | Description |
|---|
None | |
LogisticRegression | Default meta-learners are LogisticRegression for classification tasks. |
LogisticRegressionCV | Default meta-learners are LogisticRegression for classification task when CV is on. |
LightGBMClassifier | |
ElasticNet | Default meta-learners are LogisticRegression for regression task. |
ElasticNetCV | Default meta-learners are LogisticRegression for regression task when CV is on. |
LightGBMRegressor | |
LinearRegression | |
| Name | Value | Description |
|---|
| None | None | |
| LogisticRegression | LogisticRegression | Default meta-learners are LogisticRegression for classification tasks. |
| LogisticRegressionCV | LogisticRegressionCV | Default meta-learners are LogisticRegression for classification task when CV is on. |
| LightGBMClassifier | LightGBMClassifier | |
| ElasticNet | ElasticNet | Default meta-learners are LogisticRegression for regression task. |
| ElasticNetCV | ElasticNetCV | Default meta-learners are LogisticRegression for regression task when CV is on. |
| LightGBMRegressor | LightGBMRegressor | |
| LinearRegression | LinearRegression | |
| Name | Value | Description |
|---|
| None | None | |
| LogisticRegression | LogisticRegression | Default meta-learners are LogisticRegression for classification tasks. |
| LogisticRegressionCV | LogisticRegressionCV | Default meta-learners are LogisticRegression for classification task when CV is on. |
| LightGBMClassifier | LightGBMClassifier | |
| ElasticNet | ElasticNet | Default meta-learners are LogisticRegression for regression task. |
| ElasticNetCV | ElasticNetCV | Default meta-learners are LogisticRegression for regression task when CV is on. |
| LightGBMRegressor | LightGBMRegressor | |
| LinearRegression | LinearRegression | |
| Name | Value | Description |
|---|
| NONE | None | |
| LOGISTIC_REGRESSION | LogisticRegression | Default meta-learners are LogisticRegression for classification tasks. |
| LOGISTIC_REGRESSION_CV | LogisticRegressionCV | Default meta-learners are LogisticRegression for classification task when CV is on. |
| LIGHT_GBM_CLASSIFIER | LightGBMClassifier | |
| ELASTIC_NET | ElasticNet | Default meta-learners are LogisticRegression for regression task. |
| ELASTIC_NET_CV | ElasticNetCV | Default meta-learners are LogisticRegression for regression task when CV is on. |
| LIGHT_GBM_REGRESSOR | LightGBMRegressor | |
| LINEAR_REGRESSION | LinearRegression | |
| Value | Description |
|---|
None | |
LogisticRegression | Default meta-learners are LogisticRegression for classification tasks. |
LogisticRegressionCV | Default meta-learners are LogisticRegression for classification task when CV is on. |
LightGBMClassifier | |
ElasticNet | Default meta-learners are LogisticRegression for regression task. |
ElasticNetCV | Default meta-learners are LogisticRegression for regression task when CV is on. |
LightGBMRegressor | |
LinearRegression | |
Static input data definition.
- JobInputType
string | JobInputType
- [Required] Specifies the type of job.
- 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.
- Columns
map[string]string
- Mapping of column names to special uses.
- DataContext
string
- The context metadata of the data source.
- PreprocessingComponentId
string
- Reference to the component asset used to preprocess the data.
- jobInputType
String | JobInputType
- [Required] Specifies the type of job.
- 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.
- columns
Map<String,String>
- Mapping of column names to special uses.
- dataContext
String
- The context metadata of the data source.
- preprocessingComponentId
String
- Reference to the component asset used to preprocess the data.
- jobInputType
string | JobInputType
- [Required] Specifies the type of job.
- 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.
- columns
{[key: string]: string}
- Mapping of column names to special uses.
- dataContext
string
- The context metadata of the data source.
- preprocessingComponentId
string
- Reference to the component asset used to preprocess the data.
Static input data definition.
- JobInputType
string
- [Required] Specifies the type of job.
- 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.
- Columns
Dictionary<string, string>
- Mapping of column names to special uses.
- DataContext
string
- The context metadata of the data source.
- PreprocessingComponentId
string
- Reference to the component asset used to preprocess the data.
- JobInputType
string
- [Required] Specifies the type of job.
- 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.
- Columns
map[string]string
- Mapping of column names to special uses.
- DataContext
string
- The context metadata of the data source.
- PreprocessingComponentId
string
- Reference to the component asset used to preprocess the data.
- job_input_type
string
- [Required] Specifies the type of job.
- uri
string
- [Required] Input Asset URI.
- window_end
string
- [Required] The end date of the data window.
- window_start
string
- [Required] The start date of the data window.
- columns
map(string)
- Mapping of column names to special uses.
- data_context
string
- The context metadata of the data source.
- preprocessing_component_id
string
- Reference to the component asset used to preprocess the data.
- jobInputType
String
- [Required] Specifies the type of job.
- 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.
- columns
Map<String,String>
- Mapping of column names to special uses.
- dataContext
String
- The context metadata of the data source.
- preprocessingComponentId
String
- Reference to the component asset used to preprocess the data.
- jobInputType
string
- [Required] Specifies the type of job.
- 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.
- columns
{[key: string]: string}
- Mapping of column names to special uses.
- dataContext
string
- The context metadata of the data source.
- preprocessingComponentId
string
- Reference to the component asset used to preprocess the data.
- jobInputType
String
- [Required] Specifies the type of job.
- 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.
- columns
Map<String>
- Mapping of column names to special uses.
- dataContext
String
- The context metadata of the data source.
- preprocessingComponentId
String
- Reference to the component asset used to preprocess the data.
StochasticOptimizer
Stochastic optimizer for image models.
| Name | Value | Description |
|---|
| None | None | No optimizer selected. |
| Sgd | Sgd | Stochastic Gradient Descent optimizer. |
| Adam | Adam | Adam is algorithm the optimizes stochastic objective functions based on adaptive estimates of moments |
| Adamw | Adamw | AdamW is a variant of the optimizer Adam that has an improved implementation of weight decay. |
| Name | Value | Description |
|---|
| StochasticOptimizerNone | None | No optimizer selected. |
| StochasticOptimizerSgd | Sgd | Stochastic Gradient Descent optimizer. |
| StochasticOptimizerAdam | Adam | Adam is algorithm the optimizes stochastic objective functions based on adaptive estimates of moments |
| StochasticOptimizerAdamw | Adamw | AdamW is a variant of the optimizer Adam that has an improved implementation of weight decay. |
| Value | Description |
|---|
None | No optimizer selected. |
Sgd | Stochastic Gradient Descent optimizer. |
Adam | Adam is algorithm the optimizes stochastic objective functions based on adaptive estimates of moments |
Adamw | AdamW is a variant of the optimizer Adam that has an improved implementation of weight decay. |
| Name | Value | Description |
|---|
| None | None | No optimizer selected. |
| Sgd | Sgd | Stochastic Gradient Descent optimizer. |
| Adam | Adam | Adam is algorithm the optimizes stochastic objective functions based on adaptive estimates of moments |
| Adamw | Adamw | AdamW is a variant of the optimizer Adam that has an improved implementation of weight decay. |
| Name | Value | Description |
|---|
| None | None | No optimizer selected. |
| Sgd | Sgd | Stochastic Gradient Descent optimizer. |
| Adam | Adam | Adam is algorithm the optimizes stochastic objective functions based on adaptive estimates of moments |
| Adamw | Adamw | AdamW is a variant of the optimizer Adam that has an improved implementation of weight decay. |
| Name | Value | Description |
|---|
| NONE | None | No optimizer selected. |
| SGD | Sgd | Stochastic Gradient Descent optimizer. |
| ADAM | Adam | Adam is algorithm the optimizes stochastic objective functions based on adaptive estimates of moments |
| ADAMW | Adamw | AdamW is a variant of the optimizer Adam that has an improved implementation of weight decay. |
| Value | Description |
|---|
None | No optimizer selected. |
Sgd | Stochastic Gradient Descent optimizer. |
Adam | Adam is algorithm the optimizes stochastic objective functions based on adaptive estimates of moments |
Adamw | AdamW is a variant of the optimizer Adam that has an improved implementation of weight decay. |
SweepJob, SweepJobArgs
Sweep job definition.
- Objective
Pulumi.AzureNative.MachineLearningServices.Inputs.Objective
- [Required] Optimization objective.
- SamplingAlgorithm
Pulumi.AzureNative.MachineLearningServices.Inputs.BayesianSamplingAlgorithm | Pulumi.AzureNative.MachineLearningServices.Inputs.GridSamplingAlgorithm | Pulumi.AzureNative.MachineLearningServices.Inputs.RandomSamplingAlgorithm
- [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
- Trial
Pulumi.AzureNative.MachineLearningServices.Inputs.TrialComponent
- [Required] Trial component definition.
- 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
Pulumi.AzureNative.MachineLearningServices.Inputs.BanditPolicy | Pulumi.AzureNative.MachineLearningServices.Inputs.MedianStoppingPolicy | Pulumi.AzureNative.MachineLearningServices.Inputs.TruncationSelectionPolicy
- Early termination policies enable canceling poor-performing runs before they complete
- ExperimentName
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- Identity
Pulumi.AzureNative.MachineLearningServices.Inputs.AmlToken | Pulumi.AzureNative.MachineLearningServices.Inputs.ManagedIdentity | Pulumi.AzureNative.MachineLearningServices.Inputs.UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- Inputs
Dictionary<string, object>
- Mapping of input data bindings used in the job.
- IsArchived
bool
- Is the asset archived?
- Limits
Pulumi.AzureNative.MachineLearningServices.Inputs.SweepJobLimits
- Sweep Job limit.
- NotificationSetting
Pulumi.AzureNative.MachineLearningServices.Inputs.NotificationSetting
- Notification setting for the job
- Outputs
Dictionary<string, object>
- Mapping of output data bindings used in the job.
- Properties
Dictionary<string, string>
- The asset property dictionary.
- QueueSettings
Pulumi.AzureNative.MachineLearningServices.Inputs.QueueSettings
- Queue settings for the job
- Services
Dictionary<string, Pulumi.AzureNative.MachineLearningServices.Inputs.JobService>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- Tags
Dictionary<string, string>
- Tag dictionary. Tags can be added, removed, and updated.
- objective
object
- [Required] Optimization objective.
- sampling_algorithm
object | object | object
- [Required] The hyperparameter sampling algorithm
- search_space
any
- [Required] A dictionary containing each parameter and its distribution. The dictionary key is the name of the parameter
- trial
object
- [Required] Trial component definition.
- component_id
string
- ARM resource ID of the component resource.
- compute_id
string
- ARM resource ID of the compute resource.
- description
string
- The asset description text.
- display_name
string
- Display name of job.
- early_termination
object | object | object
- Early termination policies enable canceling poor-performing runs before they complete
- experiment_name
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
object | object | object
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
map(object | object | object | object | object | object | object)
- Mapping of input data bindings used in the job.
- is_archived
bool
- Is the asset archived?
- limits
object
- Sweep Job limit.
- notification_setting
object
- Notification setting for the job
- outputs
map(object | object | object | object | object | object)
- Mapping of output data bindings used in the job.
- properties
map(string)
- The asset property dictionary.
- queue_settings
object
- Queue settings for the job
- services
map(object)
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
map(string)
- Tag dictionary. Tags can be added, removed, and updated.
- objective
Objective
- [Required] Optimization objective.
- samplingAlgorithm
BayesianSamplingAlgorithm | GridSamplingAlgorithm | RandomSamplingAlgorithm
- [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
- trial
TrialComponent
- [Required] Trial component definition.
- 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
BanditPolicy | MedianStoppingPolicy | TruncationSelectionPolicy
- Early termination policies enable canceling poor-performing runs before they complete
- experimentName
String
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlToken | ManagedIdentity | UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Map<String,Object>
- Mapping of input data bindings used in the job.
- isArchived
Boolean
- Is the asset archived?
- limits
SweepJobLimits
- Sweep Job limit.
- notificationSetting
NotificationSetting
- Notification setting for the job
- outputs
Map<String,Object>
- Mapping of output data bindings used in the job.
- properties
Map<String,String>
- The asset property dictionary.
- queueSettings
QueueSettings
- Queue settings for the job
- services
Map<String,JobService>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Map<String,String>
- Tag dictionary. Tags can be added, removed, and updated.
- objective
Objective
- [Required] Optimization objective.
- sampling_algorithm
BayesianSamplingAlgorithm | GridSamplingAlgorithm | RandomSamplingAlgorithm
- [Required] The hyperparameter sampling algorithm
- search_space
Any
- [Required] A dictionary containing each parameter and its distribution. The dictionary key is the name of the parameter
- trial
TrialComponent
- [Required] Trial component definition.
- component_id
str
- ARM resource ID of the component resource.
- compute_id
str
- ARM resource ID of the compute resource.
- description
str
- The asset description text.
- display_name
str
- Display name of job.
- early_termination
BanditPolicy | MedianStoppingPolicy | TruncationSelectionPolicy
- Early termination policies enable canceling poor-performing runs before they complete
- experiment_name
str
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlToken | ManagedIdentity | UserIdentity
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Mapping[str, Union[CustomModelJobInput, LiteralJobInput, MLFlowModelJobInput, MLTableJobInput, TritonModelJobInput, UriFileJobInput, UriFolderJobInput]]
- Mapping of input data bindings used in the job.
- is_archived
bool
- Is the asset archived?
- limits
SweepJobLimits
- Sweep Job limit.
- notification_setting
NotificationSetting
- Notification setting for the job
- outputs
Mapping[str, Union[CustomModelJobOutput, MLFlowModelJobOutput, MLTableJobOutput, TritonModelJobOutput, UriFileJobOutput, UriFolderJobOutput]]
- Mapping of output data bindings used in the job.
- properties
Mapping[str, str]
- The asset property dictionary.
- queue_settings
QueueSettings
- Queue settings for the job
- services
Mapping[str, JobService]
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Mapping[str, str]
- Tag dictionary. Tags can be added, removed, and updated.
SweepJobLimits, SweepJobLimitsArgs
Sweep Job limit class.
- MaxConcurrentTrials
int
- Sweep Job max concurrent trials.
- MaxTotalTrials
int
- 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.
- MaxConcurrentTrials
int
- Sweep Job max concurrent trials.
- MaxTotalTrials
int
- 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.
- max_concurrent_trials
number
- Sweep Job max concurrent trials.
- max_total_trials
number
- 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.
- trial_timeout
string
- Sweep Job Trial timeout value.
- 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.
- maxConcurrentTrials
number
- Sweep Job max concurrent trials.
- maxTotalTrials
number
- 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.
- max_concurrent_trials
int
- Sweep Job max concurrent trials.
- max_total_trials
int
- Sweep Job max total trials.
- timeout
str
- The max run duration in ISO 8601 format, after which the job will be cancelled. Only supports duration with precision as low as Seconds.
- trial_timeout
str
- Sweep Job Trial timeout value.
- maxConcurrentTrials
Number
- Sweep Job max concurrent trials.
- maxTotalTrials
Number
- 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.
SweepJobLimitsResponse
Sweep Job limit class.
- MaxConcurrentTrials
int
- Sweep Job max concurrent trials.
- MaxTotalTrials
int
- 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.
- MaxConcurrentTrials
int
- Sweep Job max concurrent trials.
- MaxTotalTrials
int
- 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.
- max_concurrent_trials
number
- Sweep Job max concurrent trials.
- max_total_trials
number
- 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.
- trial_timeout
string
- Sweep Job Trial timeout value.
- 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.
- maxConcurrentTrials
number
- Sweep Job max concurrent trials.
- maxTotalTrials
number
- 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.
- max_concurrent_trials
int
- Sweep Job max concurrent trials.
- max_total_trials
int
- Sweep Job max total trials.
- timeout
str
- The max run duration in ISO 8601 format, after which the job will be cancelled. Only supports duration with precision as low as Seconds.
- trial_timeout
str
- Sweep Job Trial timeout value.
- maxConcurrentTrials
Number
- Sweep Job max concurrent trials.
- maxTotalTrials
Number
- 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.
SweepJobResponse
Sweep job definition.
- Objective
Pulumi.AzureNative.MachineLearningServices.Inputs.ObjectiveResponse
- [Required] Optimization objective.
- SamplingAlgorithm
Pulumi.AzureNative.MachineLearningServices.Inputs.BayesianSamplingAlgorithmResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.GridSamplingAlgorithmResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.RandomSamplingAlgorithmResponse
- [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
- Status
string
- Status of the job.
- Trial
Pulumi.AzureNative.MachineLearningServices.Inputs.TrialComponentResponse
- [Required] Trial component definition.
- 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
Pulumi.AzureNative.MachineLearningServices.Inputs.BanditPolicyResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.MedianStoppingPolicyResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.TruncationSelectionPolicyResponse
- Early termination policies enable canceling poor-performing runs before they complete
- ExperimentName
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- Identity
Pulumi.AzureNative.MachineLearningServices.Inputs.AmlTokenResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.ManagedIdentityResponse | Pulumi.AzureNative.MachineLearningServices.Inputs.UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- Inputs
Dictionary<string, object>
- Mapping of input data bindings used in the job.
- IsArchived
bool
- Is the asset archived?
- Limits
Pulumi.AzureNative.MachineLearningServices.Inputs.SweepJobLimitsResponse
- Sweep Job limit.
- NotificationSetting
Pulumi.AzureNative.MachineLearningServices.Inputs.NotificationSettingResponse
- Notification setting for the job
- Outputs
Dictionary<string, object>
- Mapping of output data bindings used in the job.
- Properties
Dictionary<string, string>
- The asset property dictionary.
- QueueSettings
Pulumi.AzureNative.MachineLearningServices.Inputs.QueueSettingsResponse
- Queue settings for the job
- Services
Dictionary<string, Pulumi.AzureNative.MachineLearningServices.Inputs.JobServiceResponse>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- Tags
Dictionary<string, string>
- Tag dictionary. Tags can be added, removed, and updated.
- objective
object
- [Required] Optimization objective.
- sampling_algorithm
object | object | object
- [Required] The hyperparameter sampling algorithm
- search_space
any
- [Required] A dictionary containing each parameter and its distribution. The dictionary key is the name of the parameter
- status
string
- Status of the job.
- trial
object
- [Required] Trial component definition.
- component_id
string
- ARM resource ID of the component resource.
- compute_id
string
- ARM resource ID of the compute resource.
- description
string
- The asset description text.
- display_name
string
- Display name of job.
- early_termination
object | object | object
- Early termination policies enable canceling poor-performing runs before they complete
- experiment_name
string
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
object | object | object
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
map(object | object | object | object | object | object | object)
- Mapping of input data bindings used in the job.
- is_archived
bool
- Is the asset archived?
- limits
object
- Sweep Job limit.
- notification_setting
object
- Notification setting for the job
- outputs
map(object | object | object | object | object | object)
- Mapping of output data bindings used in the job.
- properties
map(string)
- The asset property dictionary.
- queue_settings
object
- Queue settings for the job
- services
map(object)
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
map(string)
- Tag dictionary. Tags can be added, removed, and updated.
- objective
ObjectiveResponse
- [Required] Optimization objective.
- samplingAlgorithm
BayesianSamplingAlgorithmResponse | GridSamplingAlgorithmResponse | RandomSamplingAlgorithmResponse
- [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
- status
String
- Status of the job.
- trial
TrialComponentResponse
- [Required] Trial component definition.
- 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
BanditPolicyResponse | MedianStoppingPolicyResponse | TruncationSelectionPolicyResponse
- Early termination policies enable canceling poor-performing runs before they complete
- experimentName
String
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlTokenResponse | ManagedIdentityResponse | UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Map<String,Object>
- Mapping of input data bindings used in the job.
- isArchived
Boolean
- Is the asset archived?
- limits
SweepJobLimitsResponse
- Sweep Job limit.
- notificationSetting
NotificationSettingResponse
- Notification setting for the job
- outputs
Map<String,Object>
- Mapping of output data bindings used in the job.
- properties
Map<String,String>
- The asset property dictionary.
- queueSettings
QueueSettingsResponse
- Queue settings for the job
- services
Map<String,JobServiceResponse>
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Map<String,String>
- Tag dictionary. Tags can be added, removed, and updated.
- objective
ObjectiveResponse
- [Required] Optimization objective.
- sampling_algorithm
BayesianSamplingAlgorithmResponse | GridSamplingAlgorithmResponse | RandomSamplingAlgorithmResponse
- [Required] The hyperparameter sampling algorithm
- search_space
Any
- [Required] A dictionary containing each parameter and its distribution. The dictionary key is the name of the parameter
- status
str
- Status of the job.
- trial
TrialComponentResponse
- [Required] Trial component definition.
- component_id
str
- ARM resource ID of the component resource.
- compute_id
str
- ARM resource ID of the compute resource.
- description
str
- The asset description text.
- display_name
str
- Display name of job.
- early_termination
BanditPolicyResponse | MedianStoppingPolicyResponse | TruncationSelectionPolicyResponse
- Early termination policies enable canceling poor-performing runs before they complete
- experiment_name
str
- The name of the experiment the job belongs to. If not set, the job is placed in the "Default" experiment.
- identity
AmlTokenResponse | ManagedIdentityResponse | UserIdentityResponse
- Identity configuration. If set, this should be one of AmlToken, ManagedIdentity, UserIdentity or null.
Defaults to AmlToken if null.
- inputs
Mapping[str, Union[CustomModelJobInputResponse, LiteralJobInputResponse, MLFlowModelJobInputResponse, MLTableJobInputResponse, TritonModelJobInputResponse, UriFileJobInputResponse, UriFolderJobInputResponse]]
- Mapping of input data bindings used in the job.
- is_archived
bool
- Is the asset archived?
- limits
SweepJobLimitsResponse
- Sweep Job limit.
- notification_setting
NotificationSettingResponse
- Notification setting for the job
- outputs
Mapping[str, Union[CustomModelJobOutputResponse, MLFlowModelJobOutputResponse, MLTableJobOutputResponse, TritonModelJobOutputResponse, UriFileJobOutputResponse, UriFolderJobOutputResponse]]
- Mapping of output data bindings used in the job.
- properties
Mapping[str, str]
- The asset property dictionary.
- queue_settings
QueueSettingsResponse
- Queue settings for the job
- services
Mapping[str, JobServiceResponse]
- List of JobEndpoints.
For local jobs, a job endpoint will have an endpoint value of FileStreamObject.
- tags
Mapping[str, str]
- Tag dictionary. Tags can be added, removed, and updated.
SystemDataResponse
Metadata pertaining to creation and last modification of the resource.
- CreatedAt
string
- The timestamp of resource creation (UTC).
- CreatedBy
string
- The identity that created the resource.
- CreatedByType
string
- 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
string
- The type of identity that last modified the resource.
- CreatedAt
string
- The timestamp of resource creation (UTC).
- CreatedBy
string
- The identity that created the resource.
- CreatedByType
string
- 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
string
- The type of identity that last modified the resource.
- createdAt
String
- The timestamp of resource creation (UTC).
- createdBy
String
- The identity that created the resource.
- createdByType
String
- 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
String
- The type of identity that last modified the resource.
- createdAt
string
- The timestamp of resource creation (UTC).
- createdBy
string
- The identity that created the resource.
- createdByType
string
- 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
string
- The type of identity that last modified the resource.
- createdAt
String
- The timestamp of resource creation (UTC).
- createdBy
String
- The identity that created the resource.
- createdByType
String
- 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
String
- The type of identity that last modified the resource.
TableVerticalFeaturizationSettings, TableVerticalFeaturizationSettingsArgs
Featurization Configuration.
- BlockedTransformers
List<Union<string, Pulumi.AzureNative.MachineLearningServices.BlockedTransformers>>
- These transformers shall not be used in featurization.
- ColumnNameAndTypes
Dictionary<string, string>
- Dictionary of column name and its type (int, float, string, datetime etc).
- DatasetLanguage
string
- Dataset language, useful for the text data.
- EnableDnnFeaturization
bool
- Determines whether to use Dnn based featurizers for data featurization.
- Mode
string | Pulumi.AzureNative.MachineLearningServices.FeaturizationMode
- 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
Dictionary<string, ImmutableArray<Pulumi.AzureNative.MachineLearningServices.Inputs.ColumnTransformer>>
- User can specify additional transformers to be used along with the columns to which it would be applied and parameters for the transformer constructor.
- BlockedTransformers
[]string
- These transformers shall not be used in featurization.
- ColumnNameAndTypes
map[string]string
- Dictionary of column name and its type (int, float, string, datetime etc).
- DatasetLanguage
string
- Dataset language, useful for the text data.
- EnableDnnFeaturization
bool
- Determines whether to use Dnn based featurizers for data featurization.
- Mode
string | FeaturizationMode
- 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
map[string][]ColumnTransformer
- User can specify additional transformers to be used along with the columns to which it would be applied and parameters for the transformer constructor.
- blocked_transformers
list(string | "TextTargetEncoder" | "OneHotEncoder" | "CatTargetEncoder" | "TfIdf" | "WoETargetEncoder" | "LabelEncoder" | "WordEmbedding" | "NaiveBayes" | "CountVectorizer" | "HashOneHotEncoder")
- These transformers shall not be used in featurization.
- column_name_and_types
map(string)
- Dictionary of column name and its type (int, float, string, datetime etc).
- dataset_language
string
- Dataset language, useful for the text data.
- enable_dnn_featurization
bool
- Determines whether to use Dnn based featurizers for data featurization.
- mode
string | "Auto" | "Custom" | "Off"
- 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.
- transformer_params
map(list(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.
- blockedTransformers
List<Either<String,BlockedTransformers>>
- These transformers shall not be used in featurization.
- columnNameAndTypes
Map<String,String>
- Dictionary of column name and its type (int, float, string, datetime etc).
- datasetLanguage
String
- Dataset language, useful for the text data.
- enableDnnFeaturization
Boolean
- Determines whether to use Dnn based featurizers for data featurization.
- mode
String | FeaturizationMode
- 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
Map<String,List<ColumnTransformer>>
- User can specify additional transformers to be used along with the columns to which it would be applied and parameters for the transformer constructor.
- blockedTransformers
(string | BlockedTransformers)[]
- These transformers shall not be used in featurization.
- columnNameAndTypes
{[key: string]: string}
- Dictionary of column name and its type (int, float, string, datetime etc).
- datasetLanguage
string
- Dataset language, useful for the text data.
- enableDnnFeaturization
boolean
- Determines whether to use Dnn based featurizers for data featurization.
- mode
string | FeaturizationMode
- 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
{[key: string]: ColumnTransformer[]}
- User can specify additional transformers to be used along with the columns to which it would be applied and parameters for the transformer constructor.
- blocked_transformers
Sequence[Union[str, BlockedTransformers]]
- These transformers shall not be used in featurization.
- column_name_and_types
Mapping[str, str]
- Dictionary of column name and its type (int, float, string, datetime etc).
- dataset_language
str
- Dataset language, useful for the text data.
- enable_dnn_featurization
bool
- Determines whether to use Dnn based featurizers for data featurization.
- mode
str | FeaturizationMode
- 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.
- transformer_params
Mapping[str, Sequence[ColumnTransformer]]
- User can specify additional transformers to be used along with the columns to which it would be applied and parameters for the transformer constructor.
- blockedTransformers
List<String | "TextTargetEncoder" | "OneHotEncoder" | "CatTargetEncoder" | "TfIdf" | "WoETargetEncoder" | "LabelEncoder" | "WordEmbedding" | "NaiveBayes" | "CountVectorizer" | "HashOneHotEncoder">
- These transformers shall not be used in featurization.
- columnNameAndTypes
Map<String>
- Dictionary of column name and its type (int, float, string, datetime etc).
- datasetLanguage
String
- Dataset language, useful for the text data.
- enableDnnFeaturization
Boolean
- Determines whether to use Dnn based featurizers for data featurization.
- mode
String | "Auto" | "Custom" | "Off"
- 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
Map<List<Property Map>>
- User can specify additional transformers to be used along with the columns to which it would be applied and parameters for the transformer constructor.
TableVerticalFeaturizationSettingsResponse
Featurization Configuration.
- BlockedTransformers
List<string>
- These transformers shall not be used in featurization.
- ColumnNameAndTypes
Dictionary<string, string>
- Dictionary of column name and its type (int, float, string, datetime etc).
- DatasetLanguage
string
- Dataset language, useful for the text data.
- EnableDnnFeaturization
bool
- Determines whether to use Dnn based featurizers for data featurization.
- Mode
string
- 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
Dictionary<string, ImmutableArray<Pulumi.AzureNative.MachineLearningServices.Inputs.ColumnTransformerResponse>>
- User can specify additional transformers to be used along with the columns to which it would be applied and parameters for the transformer constructor.
- BlockedTransformers
[]string
- These transformers shall not be used in featurization.
- ColumnNameAndTypes
map[string]string
- Dictionary of column name and its type (int, float, string, datetime etc).
- DatasetLanguage
string
- Dataset language, useful for the text data.
- EnableDnnFeaturization
bool
- Determines whether to use Dnn based featurizers for data featurization.
- Mode
string
- 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
map[string][]ColumnTransformerResponse
- User can specify additional transformers to be used along with the columns to which it would be applied and parameters for the transformer constructor.
- blocked_transformers
list(string)
- These transformers shall not be used in featurization.
- column_name_and_types
map(string)
- Dictionary of column name and its type (int, float, string, datetime etc).
- dataset_language
string
- Dataset language, useful for the text data.
- enable_dnn_featurization
bool
- Determines whether to use Dnn based featurizers for data featurization.
- mode
string
- 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.
- transformer_params
map(list(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.
- blockedTransformers
List<String>
- These transformers shall not be used in featurization.
- columnNameAndTypes
Map<String,String>
- Dictionary of column name and its type (int, float, string, datetime etc).
- datasetLanguage
String
- Dataset language, useful for the text data.
- enableDnnFeaturization
Boolean
- Determines whether to use Dnn based featurizers for data featurization.
- mode
String
- 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
Map<String,List<ColumnTransformerResponse>>
- User can specify additional transformers to be used along with the columns to which it would be applied and parameters for the transformer constructor.
- blockedTransformers
string[]
- These transformers shall not be used in featurization.
- columnNameAndTypes
{[key: string]: string}
- Dictionary of column name and its type (int, float, string, datetime etc).
- datasetLanguage
string
- Dataset language, useful for the text data.
- enableDnnFeaturization
boolean
- Determines whether to use Dnn based featurizers for data featurization.
- mode
string
- 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
{[key: string]: ColumnTransformerResponse[]}
- User can specify additional transformers to be used along with the columns to which it would be applied and parameters for the transformer constructor.
- blocked_transformers
Sequence[str]
- These transformers shall not be used in featurization.
- column_name_and_types
Mapping[str, str]
- Dictionary of column name and its type (int, float, string, datetime etc).
- dataset_language
str
- Dataset language, useful for the text data.
- enable_dnn_featurization
bool
- Determines whether to use Dnn based featurizers for data featurization.
- mode
str
- 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.
- transformer_params
Mapping[str, Sequence[ColumnTransformerResponse]]
- User can specify additional transformers to be used along with the columns to which it would be applied and parameters for the transformer constructor.
- blockedTransformers
List<String>
- These transformers shall not be used in featurization.
- columnNameAndTypes
Map<String>
- Dictionary of column name and its type (int, float, string, datetime etc).
- datasetLanguage
String
- Dataset language, useful for the text data.
- enableDnnFeaturization
Boolean
- Determines whether to use Dnn based featurizers for data featurization.
- mode
String
- 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
Map<List<Property Map>>
- 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, TableVerticalLimitSettingsArgs
Job execution constraints.
TableVerticalLimitSettingsResponse
Job execution constraints.
TargetAggregationFunction
Target aggregate function.
| Name | Value | Description |
|---|
| None | None | Represent no value set. |
| Sum | Sum | |
| Max | Max | |
| Min | Min | |
| Mean | Mean | |
| Name | Value | Description |
|---|
| TargetAggregationFunctionNone | None | Represent no value set. |
| TargetAggregationFunctionSum | Sum | |
| TargetAggregationFunctionMax | Max | |
| TargetAggregationFunctionMin | Min | |
| TargetAggregationFunctionMean | Mean | |
| Value | Description |
|---|
None | Represent no value set. |
Sum | |
Max | |
Min | |
Mean | |
| Name | Value | Description |
|---|
| None | None | Represent no value set. |
| Sum | Sum | |
| Max | Max | |
| Min | Min | |
| Mean | Mean | |
| Name | Value | Description |
|---|
| None | None | Represent no value set. |
| Sum | Sum | |
| Max | Max | |
| Min | Min | |
| Mean | Mean | |
| Name | Value | Description |
|---|
| NONE | None | Represent no value set. |
| SUM | Sum | |
| MAX | Max | |
| MIN | Min | |
| MEAN | Mean | |
| Value | Description |
|---|
None | Represent no value set. |
Sum | |
Max | |
Min | |
Mean | |
TensorFlow, TensorFlowArgs
TensorFlow distribution configuration.
TensorFlowResponse
TensorFlow distribution configuration.
TextClassification, TextClassificationArgs
Text Classification task in AutoML NLP vertical.
NLP - Natural Language Processing.
TextClassificationMultilabel, TextClassificationMultilabelArgs
Text Classification Multilabel task in AutoML NLP vertical.
NLP - Natural Language Processing.
TextClassificationMultilabelResponse
Text Classification Multilabel task in AutoML NLP vertical.
NLP - Natural Language Processing.
TextClassificationResponse
Text Classification task in AutoML NLP vertical.
NLP - Natural Language Processing.
TextNer, TextNerArgs
Text-NER task in AutoML NLP vertical.
NER - Named Entity Recognition.
NLP - Natural Language Processing.
TextNerResponse
Text-NER task in AutoML NLP vertical.
NER - Named Entity Recognition.
NLP - Natural Language Processing.
TopNFeaturesByAttribution, TopNFeaturesByAttributionArgs
- Top
int
- The number of top features to include.
- Top
int
- The number of top features to include.
- top
number
- The number of top features to include.
- top
Integer
- The number of top features to include.
- top
number
- The number of top features to include.
- top
int
- The number of top features to include.
- top
Number
- The number of top features to include.
TopNFeaturesByAttributionResponse
- Top
int
- The number of top features to include.
- Top
int
- The number of top features to include.
- top
number
- The number of top features to include.
- top
Integer
- The number of top features to include.
- top
number
- The number of top features to include.
- top
int
- The number of top features to include.
- top
Number
- The number of top features to include.
TrialComponent, TrialComponentArgs
Trial component definition.
- command
string
- [Required] The command to execute on startup of the job. eg. "python train.py"
- environment_id
string
- [Required] The ARM resource ID of the Environment specification for the job.
- code_id
string
- ARM resource ID of the code asset.
- distribution
object | object | object
- Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null.
- environment_variables
map(string)
- Environment variables included in the job.
- resources
object
- Compute Resource configuration for the job.
TrialComponentResponse
Trial component definition.
- command
string
- [Required] The command to execute on startup of the job. eg. "python train.py"
- environment_id
string
- [Required] The ARM resource ID of the Environment specification for the job.
- code_id
string
- ARM resource ID of the code asset.
- distribution
object | object | object
- Distribution configuration of the job. If set, this should be one of Mpi, Tensorflow, PyTorch, or null.
- environment_variables
map(string)
- Environment variables included in the job.
- resources
object
- Compute Resource configuration for the job.
- Uri
string
- [Required] Input Asset URI.
- Description
string
- Description for the input.
- Mode
string
- Enum to determine the input data delivery mode.
- Uri
string
- [Required] Input Asset URI.
- Description
string
- Description for the input.
- Mode
string
- Enum to determine the input data delivery mode.
- uri
string
- [Required] Input Asset URI.
- description
string
- Description for the input.
- mode
string
- Enum to determine the input data delivery mode.
- uri
String
- [Required] Input Asset URI.
- description
String
- Description for the input.
- mode
String
- Enum to determine the input data delivery mode.
- uri
string
- [Required] Input Asset URI.
- description
string
- Description for the input.
- mode
string
- Enum to determine the input data delivery mode.
- uri
str
- [Required] Input Asset URI.
- description
str
- Description for the input.
- mode
str
- Enum to determine the input data delivery mode.
- uri
String
- [Required] Input Asset URI.
- description
String
- Description for the input.
- mode
String
- Enum to determine the input data delivery mode.
TritonModelJobOutput, TritonModelJobOutputArgs
TritonModelJobOutputResponse
- AssetName
string
- Output Asset Name.
- Description
string
- Description for the output.
- Mode
string
- Output data delivery mode enums.
- Uri
string
- Output Asset URI.
- AssetName
string
- Output Asset Name.
- Description
string
- Description for the output.
- Mode
string
- Output data delivery mode enums.
- Uri
string
- Output Asset URI.
- asset_name
string
- Output Asset Name.
- description
string
- Description for the output.
- mode
string
- Output data delivery mode enums.
- uri
string
- Output Asset URI.
- assetName
String
- Output Asset Name.
- description
String
- Description for the output.
- mode
String
- Output data delivery mode enums.
- uri
String
- Output Asset URI.
- assetName
string
- Output Asset Name.
- description
string
- Description for the output.
- mode
string
- Output data delivery mode enums.
- uri
string
- Output Asset URI.
- assetName
String
- Output Asset Name.
- description
String
- Description for the output.
- mode
String
- Output data delivery mode enums.
- uri
String
- Output Asset URI.
TruncationSelectionPolicy, TruncationSelectionPolicyArgs
Defines an early termination policy that cancels a given percentage of runs at each evaluation interval.
TruncationSelectionPolicyResponse
Defines an early termination policy that cancels a given percentage of runs at each evaluation interval.
- Uri
string
- [Required] Input Asset URI.
- Description
string
- Description for the input.
- Mode
string
- Enum to determine the input data delivery mode.
- Uri
string
- [Required] Input Asset URI.
- Description
string
- Description for the input.
- Mode
string
- Enum to determine the input data delivery mode.
- uri
string
- [Required] Input Asset URI.
- description
string
- Description for the input.
- mode
string
- Enum to determine the input data delivery mode.
- uri
String
- [Required] Input Asset URI.
- description
String
- Description for the input.
- mode
String
- Enum to determine the input data delivery mode.
- uri
string
- [Required] Input Asset URI.
- description
string
- Description for the input.
- mode
string
- Enum to determine the input data delivery mode.
- uri
str
- [Required] Input Asset URI.
- description
str
- Description for the input.
- mode
str
- Enum to determine the input data delivery mode.
- uri
String
- [Required] Input Asset URI.
- description
String
- Description for the input.
- mode
String
- Enum to determine the input data delivery mode.
UriFileJobOutput, UriFileJobOutputArgs
UriFileJobOutputResponse
- AssetName
string
- Output Asset Name.
- Description
string
- Description for the output.
- Mode
string
- Output data delivery mode enums.
- Uri
string
- Output Asset URI.
- AssetName
string
- Output Asset Name.
- Description
string
- Description for the output.
- Mode
string
- Output data delivery mode enums.
- Uri
string
- Output Asset URI.
- asset_name
string
- Output Asset Name.
- description
string
- Description for the output.
- mode
string
- Output data delivery mode enums.
- uri
string
- Output Asset URI.
- assetName
String
- Output Asset Name.
- description
String
- Description for the output.
- mode
String
- Output data delivery mode enums.
- uri
String
- Output Asset URI.
- assetName
string
- Output Asset Name.
- description
string
- Description for the output.
- mode
string
- Output data delivery mode enums.
- uri
string
- Output Asset URI.
- assetName
String
- Output Asset Name.
- description
String
- Description for the output.
- mode
String
- Output data delivery mode enums.
- uri
String
- Output Asset URI.
- Uri
string
- [Required] Input Asset URI.
- Description
string
- Description for the input.
- Mode
string
- Enum to determine the input data delivery mode.
- Uri
string
- [Required] Input Asset URI.
- Description
string
- Description for the input.
- Mode
string
- Enum to determine the input data delivery mode.
- uri
string
- [Required] Input Asset URI.
- description
string
- Description for the input.
- mode
string
- Enum to determine the input data delivery mode.
- uri
String
- [Required] Input Asset URI.
- description
String
- Description for the input.
- mode
String
- Enum to determine the input data delivery mode.
- uri
string
- [Required] Input Asset URI.
- description
string
- Description for the input.
- mode
string
- Enum to determine the input data delivery mode.
- uri
str
- [Required] Input Asset URI.
- description
str
- Description for the input.
- mode
str
- Enum to determine the input data delivery mode.
- uri
String
- [Required] Input Asset URI.
- description
String
- Description for the input.
- mode
String
- Enum to determine the input data delivery mode.
UriFolderJobOutput, UriFolderJobOutputArgs
UriFolderJobOutputResponse
- AssetName
string
- Output Asset Name.
- Description
string
- Description for the output.
- Mode
string
- Output data delivery mode enums.
- Uri
string
- Output Asset URI.
- AssetName
string
- Output Asset Name.
- Description
string
- Description for the output.
- Mode
string
- Output data delivery mode enums.
- Uri
string
- Output Asset URI.
- asset_name
string
- Output Asset Name.
- description
string
- Description for the output.
- mode
string
- Output data delivery mode enums.
- uri
string
- Output Asset URI.
- assetName
String
- Output Asset Name.
- description
String
- Description for the output.
- mode
String
- Output data delivery mode enums.
- uri
String
- Output Asset URI.
- assetName
string
- Output Asset Name.
- description
string
- Description for the output.
- mode
string
- Output data delivery mode enums.
- uri
string
- Output Asset URI.
- assetName
String
- Output Asset Name.
- description
String
- Description for the output.
- mode
String
- Output data delivery mode enums.
- uri
String
- Output Asset URI.
UseStl
Configure STL Decomposition of the time-series target column.
| Name | Value | Description |
|---|
| None | None | No stl decomposition. |
| Season | Season | |
| SeasonTrend | SeasonTrend | |
| Name | Value | Description |
|---|
| UseStlNone | None | No stl decomposition. |
| UseStlSeason | Season | |
| UseStlSeasonTrend | SeasonTrend | |
| Value | Description |
|---|
None | No stl decomposition. |
Season | |
SeasonTrend | |
| Name | Value | Description |
|---|
| None | None | No stl decomposition. |
| Season | Season | |
| SeasonTrend | SeasonTrend | |
| Name | Value | Description |
|---|
| None | None | No stl decomposition. |
| Season | Season | |
| SeasonTrend | SeasonTrend | |
| Name | Value | Description |
|---|
| NONE | None | No stl decomposition. |
| SEASON | Season | |
| SEASON_TREND | SeasonTrend | |
| Value | Description |
|---|
None | No stl decomposition. |
Season | |
SeasonTrend | |
UserAssignedIdentityResponse
User assigned identity properties
- ClientId
string
- The client ID of the assigned identity.
- PrincipalId
string
- The principal ID of the assigned identity.
- ClientId
string
- The client ID of the assigned identity.
- PrincipalId
string
- The principal ID of the assigned identity.
- client_id
string
- The client ID of the assigned identity.
- principal_id
string
- The principal ID of the assigned identity.
- clientId
String
- The client ID of the assigned identity.
- principalId
String
- The principal ID of the assigned identity.
- clientId
string
- The client ID of the assigned identity.
- principalId
string
- The principal ID of the assigned identity.
- client_id
str
- The client ID of the assigned identity.
- principal_id
str
- The principal ID of the assigned identity.
- clientId
String
- The client ID of the assigned identity.
- principalId
String
- The principal ID of the assigned identity.
UserIdentity, UserIdentityArgs
User identity configuration.
UserIdentityResponse
User identity configuration.
ValidationMetricType
Metric computation method to use for validation metrics in image tasks.
| Name | Value | Description |
|---|
| None | None | No metric. |
| Coco | Coco | Coco metric. |
| Voc | Voc | Voc metric. |
| CocoVoc | CocoVoc | CocoVoc metric. |
| Name | Value | Description |
|---|
| ValidationMetricTypeNone | None | No metric. |
| ValidationMetricTypeCoco | Coco | Coco metric. |
| ValidationMetricTypeVoc | Voc | Voc metric. |
| ValidationMetricTypeCocoVoc | CocoVoc | CocoVoc metric. |
| Value | Description |
|---|
None | No metric. |
Coco | Coco metric. |
Voc | Voc metric. |
CocoVoc | CocoVoc metric. |
| Name | Value | Description |
|---|
| None | None | No metric. |
| Coco | Coco | Coco metric. |
| Voc | Voc | Voc metric. |
| CocoVoc | CocoVoc | CocoVoc metric. |
| Name | Value | Description |
|---|
| None | None | No metric. |
| Coco | Coco | Coco metric. |
| Voc | Voc | Voc metric. |
| CocoVoc | CocoVoc | CocoVoc metric. |
| Name | Value | Description |
|---|
| NONE | None | No metric. |
| COCO | Coco | Coco metric. |
| VOC | Voc | Voc metric. |
| COCO_VOC | CocoVoc | CocoVoc metric. |
| Value | Description |
|---|
None | No metric. |
Coco | Coco metric. |
Voc | Voc metric. |
CocoVoc | CocoVoc metric. |
WeekDay
Enum of weekday
| Name | Value | Description |
|---|
| Monday | Monday | Monday weekday |
| Tuesday | Tuesday | Tuesday weekday |
| Wednesday | Wednesday | Wednesday weekday |
| Thursday | Thursday | Thursday weekday |
| Friday | Friday | Friday weekday |
| Saturday | Saturday | Saturday weekday |
| Sunday | Sunday | Sunday weekday |
| Name | Value | Description |
|---|
| WeekDayMonday | Monday | Monday weekday |
| WeekDayTuesday | Tuesday | Tuesday weekday |
| WeekDayWednesday | Wednesday | Wednesday weekday |
| WeekDayThursday | Thursday | Thursday weekday |
| WeekDayFriday | Friday | Friday weekday |
| WeekDaySaturday | Saturday | Saturday weekday |
| WeekDaySunday | Sunday | Sunday weekday |
| Value | Description |
|---|
Monday | Monday weekday |
Tuesday | Tuesday weekday |
Wednesday | Wednesday weekday |
Thursday | Thursday weekday |
Friday | Friday weekday |
Saturday | Saturday weekday |
Sunday | Sunday weekday |
| Name | Value | Description |
|---|
| Monday | Monday | Monday weekday |
| Tuesday | Tuesday | Tuesday weekday |
| Wednesday | Wednesday | Wednesday weekday |
| Thursday | Thursday | Thursday weekday |
| Friday | Friday | Friday weekday |
| Saturday | Saturday | Saturday weekday |
| Sunday | Sunday | Sunday weekday |
| Name | Value | Description |
|---|
| Monday | Monday | Monday weekday |
| Tuesday | Tuesday | Tuesday weekday |
| Wednesday | Wednesday | Wednesday weekday |
| Thursday | Thursday | Thursday weekday |
| Friday | Friday | Friday weekday |
| Saturday | Saturday | Saturday weekday |
| Sunday | Sunday | Sunday weekday |
| Name | Value | Description |
|---|
| MONDAY | Monday | Monday weekday |
| TUESDAY | Tuesday | Tuesday weekday |
| WEDNESDAY | Wednesday | Wednesday weekday |
| THURSDAY | Thursday | Thursday weekday |
| FRIDAY | Friday | Friday weekday |
| SATURDAY | Saturday | Saturday weekday |
| SUNDAY | Sunday | Sunday weekday |
| Value | Description |
|---|
Monday | Monday weekday |
Tuesday | Tuesday weekday |
Wednesday | Wednesday weekday |
Thursday | Thursday weekday |
Friday | Friday weekday |
Saturday | Saturday weekday |
Sunday | Sunday weekday |
Import
An existing resource can be imported using its type token, name, and identifier, e.g.
$ pulumi import azure-native:machinelearningservices:Schedule string /subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.MachineLearningServices/workspaces/{workspaceName}/schedules/{name}
To learn more about importing existing cloud resources, see Importing resources.
Package Details
- Repository
- Azure Native pulumi/pulumi-azure-native
- License
- Apache-2.0