ImageSweepSettings Class

  • Reference

Sweep settings for all AutoML Image Verticals.

] :keyword early_termination: Type of early termination policy. :paramtype early_termination: Union[

~azure.mgmt.machinelearningservices.models.BanditPolicy, ~azure.mgmt.machinelearningservices.models.MedianStoppingPolicy, ~azure.mgmt.machinelearningservices.models.TruncationSelectionPolicy

]

Inheritance
azure.ai.ml.entities._mixins.RestTranslatableMixin
ImageSweepSettings

Constructor

ImageSweepSettings(*, sampling_algorithm: str | Grid | Bayesian | Random, early_termination: EarlyTerminationPolicy | BanditPolicy | MedianStoppingPolicy | TruncationSelectionPolicy | None = None)

Keyword-Only Parameters

Name Description
sampling_algorithm
Union[ str, GRID, BAYESIAN, RANDOM

Required. Type of the hyperparameter sampling. algorithms. Possible values include: "Grid", "Random", "Bayesian".
early_termination
Required

Examples

Defining the sweep settings for an automl image job.


   from azure.ai.ml import automl
   from azure.ai.ml.sweep import BanditPolicy

   image_sweep_settings = automl.ImageSweepSettings(
       sampling_algorithm="Grid",
       early_termination=BanditPolicy(evaluation_interval=2, slack_factor=0.05, delay_evaluation=6),
   )