PipelineRunOutput Class

Data class used to encapsulate return values from calling run_pipeline.

Initialize a PipelineRunOutput object.

Inheritance

builtins.object

PipelineRunOutput

Constructor

PipelineRunOutput(task_type: str, enable_streaming: bool, pipeline_obj: Pipeline, training_type: str)

Parameters

Name	Description
pipeline_obj Required	the pipeline being executed
training_type Required	the training type
task_type Required
enable_streaming Required

Methods

record_pipeline_output

Save output from a successful pipeline execution.

record_pipeline_output

Save output from a successful pipeline execution.

record_pipeline_output(scores: Dict[str, Any], fit_time: float, fitted_pipeline: Pipeline, fitted_pipelines_train: List[Pipeline] | None, training_percent: int) -> None

Parameters

Name	Description
scores Required	a dictionary containing metric scores
fit_time Required	the time taken to execute the pipeline, in seconds
fitted_pipeline Required	the fitted model
fitted_pipelines_train Required	the partially trained pipelines when using cross validation
training_percent Required	percent of data that was used for training

Attributes

fit_time

Get the fit time in seconds.

fitted_pipeline

Get the fitted model.

fitted_pipelines_train

Get the partially trained fitted models.

pretrain_props

Get the pretrain properties.

pretrain_props_sanitized

Get the pretrain properties with None converted to empty string.

run_properties

Get the pipeline run properties.

scores

Get the scores.

training_percent

Get the training percent used for this pipeline

training_type

Get the training type.