progress#

Progress abstraction holding stopping criteria and progress information.

Module Contents#

Classes Summary#

Progress

Progress object holding stopping criteria and progress information.

Contents#

class evalml.automl.progress.Progress(max_time=None, max_batches=None, max_iterations=None, patience=None, tolerance=None, automl_algorithm=None, objective=None, verbose=False)[source]#

Progress object holding stopping criteria and progress information.

Parameters

  • max_time (int) – Maximum time to search for pipelines.

  • max_iterations (int) – Maximum number of iterations to search.

  • max_batches (int) – The maximum number of batches of pipelines to search. Parameters max_time, and max_iterations have precedence over stopping the search.

  • patience (int) – Number of iterations without improvement to stop search early.

  • tolerance (float) – Minimum percentage difference to qualify as score improvement for early stopping.

  • automl_algorithm (str) – The automl algorithm to use. Used to calculate iterations if max_batches is selected as stopping criteria.

  • objective (str, ObjectiveBase) – The objective used in search.

  • verbose (boolean) – Whether or not to log out stopping information.

Methods

elapsed

Return time elapsed using the start time and current time.

return_progress

Return information about current and end state of each stopping criteria in order of priority.

should_continue

Given AutoML Results, return whether or not the search should continue.

start_timing

Sets start time to current time.
elapsed(self)[source]#

Return time elapsed using the start time and current time.

return_progress(self)[source]#

Return information about current and end state of each stopping criteria in order of priority.

Returns

list of dictionaries containing information of each stopping criteria.

Return type

List[Dict[str, unit]]

should_continue(self, results, interrupted=False, mid_batch=False)[source]#

Given AutoML Results, return whether or not the search should continue.

Parameters

  • results (dict) – AutoMLSearch results.

  • interrupted (bool) – whether AutoMLSearch was given an keyboard interrupt. Defaults to False.

  • mid_batch (bool) – whether this method was called while in the middle of a batch or not. Defaults to False.

Returns

True if search should continue, False otherwise.

Return type

bool

start_timing(self)[source]#

Sets start time to current time.