Delete Experiments and Runs

experiments_addon.delete.has_metric(metrics: List[TrialComponentMetricSummary], metric_name: str) → bool

Checks if Metric Summaries contain the searched metric summary

Parameters:

  • metrics (list of TrialComponentMetricSummary) – List of trial component metrics.

  • metric_name (str) – Metric name to look at.

Returns:

Boolean that is true if metric name is in trial component metrics summary

experiments_addon.delete.delete_run_without_metric(experiment_name: str, metric_name: str = 'val_loss', sagemaker_session: Session | None = None) → None

Delete all runs for a given experiment_name i.e. trial components that do not have the given metric stored.

Function can be useful to delete all experiment_name run that have been started but failed or have been aborted. E.g. due to error in the code or early detected problems with the model.

Parameters:

  • experiment_name (str) – Determines the experiment_name where the run to delete are.

  • metric_name (str) – Determines the metric by name to check for availability.

  • sagemaker_session (Session) – Session object which manages interactions with Amazon SageMaker APIs If not specified, one is created using the default AWS configuration chain.

experiments_addon.delete.delete_runs_like(experiment_name: str, name_substr: str, sagemaker_session: Session | None = None) → None

Delete all runs of an experiment_name that fulfill <name_substr in run_name>

Function can be useful to delete several runs with similar names from the experimentation board. The function delete all run with run names that contain the name_substr.

Parameters:

  • experiment_name (str) – Determines the experiment_name where the run to delete are.

  • name_substr (str) – If this substring is in the job name. The job will be deleted.

  • sagemaker_session (Session) – Session object which manages interactions with Amazon SageMaker APIs If not specified, one is created using the default AWS configuration chain.

experiments_addon.delete.delete_experiment(experiment_name: str, sagemaker_session: Session | None = None) → None

Delete experiment_name and associated runs.

Parameters:

  • experiment_name (str) – Determines the experiment_name to delete.

  • sagemaker_session (Session) – Session object which manages interactions with Amazon SageMaker APIs If not specified, one is created using the default AWS configuration chain.

  • experiment_name – Experiment to delete.