Callbacks¶

`base.BaseCallback`()	Base Callback from which all Callbacks must inherit from
`ProgressBar`(**kwargs)	Displays a tqdm progress bar with the training progress.
`ConsecutiveStopping`(generations[, metric])	Stop the optimization if the current metric value is no greater that at least one metric from the last N generations
`DeltaThreshold`(threshold[, generations])	Stops the optimization if the absolute difference between the maximum and minimum value from the last N generations is less or equals to a threshold.
`TimerStopping`(total_seconds)	Stops the optimization process if a limit training time has been elapsed.
`ThresholdStopping`(threshold[, metric])	Stop the optimization if the metric from cross validation score is greater or equals than the define threshold
`TensorBoard`([log_dir, run_id])	Log all the fitness metrics to Tensorboard into log_dir/run_id folder
`LogbookSaver`(checkpoint_path, **dump_options)	Saves the estimator.logbook parameter chapter object in a local file system.

class sklearn_genetic.callbacks.base.BaseCallback[source]¶

Base Callback from which all Callbacks must inherit from

on_end(logbook=None, estimator=None)[source]¶

Take actions at the end of the training

Parameters

logbook:: Current stream logbook with the stats required
estimator:: GASearchCV Estimator that is being optimized

on_start(estimator=None)[source]¶

Take actions at the start of the training

Parameters

estimator:: GASearchCV Estimator that is being optimized

on_step(record=None, logbook=None, estimator=None)[source]¶

Take actions after fitting each generation.

Parameters

record: dict: default=None: A logbook record
logbook:: Current stream logbook with the stats required
estimator:: GASearchCV Estimator that is being optimized

Returns

decision: bool, default=False: If True, the optimization process is stopped, else, if continues to the next generation.

class sklearn_genetic.callbacks.ProgressBar(**kwargs)[source]¶

Displays a tqdm progress bar with the training progress.

Parameters

kwargs: dict, default = {“file”: sys.stdout}: A dict with valid arguments from tqdm.auto.tqdm

on_end(logbook=None, estimator=None)[source]¶: Closes the progress bar

on_start(estimator=None)[source]¶: Initializes the progress bar with the kwargs and total generations

on_step(record=None, logbook=None, estimator=None)[source]¶: Increases the progress bar by one step

class sklearn_genetic.callbacks.ConsecutiveStopping(generations, metric='fitness')[source]¶

Stop the optimization if the current metric value is no greater that at least one metric from the last N generations

Parameters

generations: int, default=None: Number of current generations to compare against current generation
metric: {‘fitness’, ‘fitness_std’, ‘fitness_max’, ‘fitness_min’}, default =’fitness’: Name of the metric inside ‘record’ logged in each iteration

on_step(record=None, logbook=None, estimator=None)[source]¶

Take actions after fitting each generation.

Parameters

record: dict: default=None: A logbook record
logbook:: Current stream logbook with the stats required
estimator:: GASearchCV Estimator that is being optimized

Returns

decision: bool, default=False: If True, the optimization process is stopped, else, if continues to the next generation.

class sklearn_genetic.callbacks.DeltaThreshold(threshold, generations=2, metric: str = 'fitness')[source]¶

Stops the optimization if the absolute difference between the maximum and minimum value from the last N generations is less or equals to a threshold. The threshold gets evaluated after the number of generations specified is reached.

Parameters

threshold: float, default=None: Threshold to compare the differences between cross-validation scores.
generations: int, default=2: Number of generations to compare, includes the current generation.
metric: {‘fitness’, ‘fitness_std’, ‘fitness_max’, ‘fitness_min’}, default =’fitness’: Name of the metric inside ‘record’ logged in each iteration.

on_step(record=None, logbook=None, estimator=None)[source]¶

Take actions after fitting each generation.

Parameters

record: dict: default=None: A logbook record
logbook:: Current stream logbook with the stats required
estimator:: GASearchCV Estimator that is being optimized

Returns

decision: bool, default=False: If True, the optimization process is stopped, else, if continues to the next generation.

class sklearn_genetic.callbacks.TimerStopping(total_seconds)[source]¶

Stops the optimization process if a limit training time has been elapsed. This time is checked after each generation fit

Parameters

total_seconds: int: Total time in seconds that the estimator is allowed to fit

on_start(estimator=None)[source]¶

Take actions at the start of the training

Parameters

estimator:: GASearchCV Estimator that is being optimized

on_step(record=None, logbook=None, estimator=None)[source]¶

Take actions after fitting each generation.

Parameters

record: dict: default=None: A logbook record
logbook:: Current stream logbook with the stats required
estimator:: GASearchCV Estimator that is being optimized

Returns

decision: bool, default=False: If True, the optimization process is stopped, else, if continues to the next generation.

class sklearn_genetic.callbacks.ThresholdStopping(threshold, metric='fitness')[source]¶

Stop the optimization if the metric from cross validation score is greater or equals than the define threshold

Parameters

threshold: float, default=None: Threshold to compare against the current cross validation average score and determine if the optimization process must stop
metric: {‘fitness’, ‘fitness_std’, ‘fitness_max’, ‘fitness_min’}, default =’fitness’: Name of the metric inside ‘record’ logged in each iteration

on_step(record=None, logbook=None, estimator=None)[source]¶

Take actions after fitting each generation.

Parameters

record: dict: default=None: A logbook record
logbook:: Current stream logbook with the stats required
estimator:: GASearchCV Estimator that is being optimized

Returns

decision: bool, default=False: If True, the optimization process is stopped, else, if continues to the next generation.

class sklearn_genetic.callbacks.TensorBoard(log_dir='./logs', run_id=None)[source]¶

Log all the fitness metrics to Tensorboard into log_dir/run_id folder

Parameters

log_dir: str, default=”./logs”: Path to the main folder where the data will be log
run_id: str, default=None: Subfolder where the data will be log, if None it will create a folder with the current datetime with format time.strftime(“%Y_%m_%d-%H_%M_%S”)

on_step(record=None, logbook=None, estimator=None)[source]¶

Take actions after fitting each generation.

Parameters

record: dict: default=None: A logbook record
logbook:: Current stream logbook with the stats required
estimator:: GASearchCV Estimator that is being optimized

Returns

decision: bool, default=False: If True, the optimization process is stopped, else, if continues to the next generation.

class sklearn_genetic.callbacks.LogbookSaver(checkpoint_path, **dump_options)[source]¶

Saves the estimator.logbook parameter chapter object in a local file system.

Parameters

checkpoint_path: str: Location where checkpoint will be saved to
dump_options, str: Valid kwargs from joblib dump

on_step(record=None, logbook=None, estimator=None)[source]¶

Take actions after fitting each generation.

Parameters

record: dict: default=None: A logbook record
logbook:: Current stream logbook with the stats required
estimator:: GASearchCV Estimator that is being optimized

Returns

decision: bool, default=False: If True, the optimization process is stopped, else, if continues to the next generation.