neuraxle.plotting

Module-level documentation for neuraxle.plotting. Here is an inheritance diagram, including dependencies to other base modules of Neuraxle:

Inheritance diagram of neuraxle.plotting

Notebook matplotlib plotting functions

Utility function for plotting in notebooks.

Functions

plot_distribution_space(hyperparameter_space)

plot_histogram(title, distribution[, num_bins])

plot_pdf_cdf(title, distribution)

Classes

TrialMetricsPlottingObserver(…)

An observer that receives trial updates and plots metric results.


neuraxle.plotting.plot_histogram(title: str, distribution: neuraxle.hyperparams.distributions.HyperparameterDistribution, num_bins=50)[source]
neuraxle.plotting.plot_pdf_cdf(title: str, distribution: neuraxle.hyperparams.distributions.HyperparameterDistribution)[source]
neuraxle.plotting.plot_distribution_space(hyperparameter_space: neuraxle.hyperparams.space.HyperparameterSpace, num_bins=50)[source]
class neuraxle.plotting.TrialMetricsPlottingObserver(plotting_folder_name: str = 'metric_results', save_plots: bool = True, plot_trial_on_next: bool = True, plot_all_trials_on_complete: bool = True, plot_individual_trials_on_complete: bool = True)[source]

Bases: neuraxle.metaopt.observable._Observer

An observer that receives trial updates and plots metric results. It can plot individual trials on each update, or upon completion. It can also plot all trials in the same plot upon completion.

Usage Example:

hyperparams_repository: HyperparamsJSONRepository = HyperparamsJSONRepository(cache_folder='trials')
hyperparams_repository.subscribe(TrialMetricsPlottingObserver(
    plotting_folder_name: str = 'metric_results',
    plot_individual_trials_on_complete=False,
    plot_trial_on_next=True,
    plot_all_trials_on_complete=False,
    save_plots=True
))

auto_ml = AutoML(
    pipeline,
    n_trials=n_iter,
    validation_split_function=validation_splitter(0.2),
    hyperparams_optimizer=RandomSearchHyperparameterSelectionStrategy(),
    scoring_callback=ScoringCallback(mean_squared_error, higher_score_is_better=False),
    callbacks=[
        MetricCallback('mse', metric_function=mean_squared_error, higher_score_is_better=False)
    ],
    refit_trial=True,
    cache_folder_when_no_handle=str(tmpdir)
)

auto_ml = auto_ml.fit(data_inputs, expected_outputs)

See also

_Observer, Trial, Trials, AutoML, HyperparamsRepository, HyperparamsJSONRepository

__init__(plotting_folder_name: str = 'metric_results', save_plots: bool = True, plot_trial_on_next: bool = True, plot_all_trials_on_complete: bool = True, plot_individual_trials_on_complete: bool = True)[source]

Initialize self. See help(type(self)) for accurate signature.

on_next(value: Tuple[neuraxle.metaopt.auto_ml.HyperparamsRepository, neuraxle.metaopt.trial.Trial])[source]

Plot updated trial metric results.

Parameters

value – hyperparams_repository, trial

Returns

_plot_all_trial_main_and_validation_metric_results(repo, trial)[source]
on_complete(value: neuraxle.metaopt.auto_ml.HyperparamsRepository)[source]

Plot trial metric results upon completion.

Parameters

value (HyperparamsRepository) – hyperparams_repository, trial

Returns

_plot_all_trials_on_complete(repo, trials)[source]
_plot_all_trials_validation_results_for_metric(trials, metric_name, cache_folder, split_number)[source]
_plot_all_trials_training_results_for_metric(trials, metric_name, cache_folder, split_number)[source]
_show_or_save_plot(plotting_file)[source]