AutoMLImageConfig Class

  • Reference

Represents configuration for submitting an automated ML image experiment in Azure Machine Learning.

This configuration object contains and persists the parameters for configuring the experiment run, as well as the training data to be used at run time. For guidance on selecting your settings, see: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models.

Create an AutoMLImageConfig.

Inheritance
AutoMLConfig
AutoMLImageConfig

Constructor

AutoMLImageConfig(task: ImageTask, compute_target: Any, training_data: TabularDataset, hyperparameter_sampling: HyperParameterSampling, iterations: int, max_concurrent_iterations: int | None = None, experiment_timeout_hours: float | int | None = None, early_termination_policy: EarlyTerminationPolicy | None = None, validation_data: TabularDataset | None = None, arguments: List[Any] | None = None, **kwargs: Any)

Parameters

task
<xref:ImageTask>
Required

The type of task to run.

compute_target
Any
Required

The Azure Machine Learning compute target to run the ML image experiment on. Only remote GPU computes with more than 12 GB of GPU memory are supported. See https://docs.microsoft.com/azure/machine-learning/how-to-auto-train-remote for more information on compute targets.

training_data
<xref:TabularDataset>
Required

The training data to be used within the experiment.

hyperparameter_sampling
<xref:HyperParameterSampling>
Required

Object containing the hyperparameter space, the sampling method, and in some cases additional properties for specific sampling classes.

iterations
int
Required

The total number of different model and parameter combinations to test during an automated ML image experiment. If not specified, the default is 1 iteration.

max_concurrent_iterations
Optional[int]
default value: None

Represents the maximum number of iterations that would be executed in parallel. The default value is the same as the number of iterations provided.

experiment_timeout_hours
Optional[Union[float, int]]
default value: None

Maximum amount of time in hours that all iterations combined can take before the experiment terminates. Can be a decimal value like 0.25 representing 15 minutes. If not specified, the default experiment timeout is 6 days.

early_termination_policy
Optional[<xref:EarlyTerminationPolicy>]
default value: None

Early termination policy use when using hyperparameter tuning with several iterations. An iteration is cancelled when the criteria of a specified policy are met.

validation_data
Optional[<xref:TabularDataset>]
default value: None

The validation data to be used within the experiment.

arguments
Optional[List[Any]]
default value: None

Arguments to be passed to the remote script runs. Arguments are passed in name-value pairs and the name must be prefixed by a double dash.

task
<xref:ImageTask>
Required

The type of task to run.

compute_target
Any
Required

The Azure Machine Learning compute target to run the ML image experiment on. Only remote GPU computes with more than 12 GB of GPU memory are supported. See https://docs.microsoft.com/azure/machine-learning/how-to-auto-train-remote for more information on compute targets.

training_data
<xref:TabularDataset>
Required

The training data to be used within the experiment.

hyperparameter_sampling
<xref:HyperParameterSampling>
Required

Object containing the hyperparameter space, the sampling method, and in some cases additional properties for specific sampling classes.

iterations
int
Required

The total number of different model and parameter combinations to test during an automated ML image experiment. If not specified, the default is 1 iteration.

max_concurrent_iterations
Optional[int]
Required

Represents the maximum number of iterations that would be executed in parallel. The default value is the same as the number of iterations provided.

experiment_timeout_hours
Optional[Union[float, int]]
Required

Maximum amount of time in hours that all iterations combined can take before the experiment terminates. Can be a decimal value like 0.25 representing 15 minutes. If not specified, the default experiment timeout is 6 days.

early_termination_policy
Optional[<xref:EarlyTerminationPolicy>]
Required

Early termination policy use when using hyperparameter tuning with several iterations. An iteration is cancelled when the criteria of a specified policy are met.

validation_data
Optional[<xref:TabularDataset>]
Required

The validation data to be used within the experiment.

arguments
Optional[List[Any]]
Required

Arguments to be passed to the remote script runs. Arguments are passed in name-value pairs and the name must be prefixed by a double dash.