enchanter.callbacks

A set of functions to be executed during training, verification and testing.

Callback

Callback

class enchanter.callbacks.Callback

Bases: abc.ABC

on_epoch_end(epoch, logs: Optional[Dict] = None)

Called when the epoch ends.

on_epoch_start(epoch, logs: Optional[Dict] = None)

Called when the epoch begins.

on_test_end(logs: Optional[Dict] = None)

Called when the test ends.

on_test_start(logs: Optional[Dict] = None)

Called when the test begins.

on_test_step_end(logs: Optional[Dict] = None)

Called when the test batch ends. You can access the output of test_step.

on_test_step_start(logs: Optional[Dict] = None)

on_train_end(logs: Optional[Dict] = None)

Called when the train ends.

on_train_start(logs: Optional[Dict] = None)

Called when the train begins.

on_train_step_end(logs: Optional[Dict] = None)

Called when the training batch ends. You can access the output of train_step.

on_train_step_start(logs: Optional[Dict] = None)

Called when the training batch begins.

on_validation_end(logs: Optional[Dict] = None)

Called when the validation loop ends.

on_validation_start(logs: Optional[Dict] = None)

Called when the validation loop begins.

on_validation_step_end(logs: Optional[Dict] = None)

Called when the validation batch ends. You can access the output of val_step.

on_validation_step_start(logs: Optional[Dict] = None)

set_device(device)

set_model(model)

set_optimizer(optimizer)

CallbackManager

CallbackManager

class enchanter.callbacks.CallbackManager(callbacks: Optional[List[enchanter.callbacks.base.Callback]] = None)

A class for managing Callback.

Examples

>>> from enchanter.callbacks import EarlyStopping
>>> manager = CallbackManager([
>>>     EarlyStopping()
>>> ])
>>> for epoch in range(5):
>>>     logs = ...
>>>     manager.on_epoch_end(epoch, logs)
>>>     if manager.stop_runner:
>>>         break

EarlyStopping

EarlyStopping

class enchanter.callbacks.EarlyStopping(monitor: str = 'val_avg_loss', min_delta: float = 0.0, patience: int = 0, mode: str = 'auto')

Bases: enchanter.callbacks.base.Callback

The training ends when the value to be monitored stops changing.

Examples

>>> from enchanter.tasks import ClassificationRunner
>>> runner = ClassificationRunner(callbacks=[ClassificationRunner()])

Initializer

Parameters

• monitor – You can choose the return values of train_end(), val_end(), or test_end(). To specify, in the case of the return value of train_end(), train_XXX, in the case of the return value of val_end(), val_XXX, in the case of the return value of test_end(), test_XXX.

• min_delta – Minimum value of change determined as improvement for the monitored value.

• patience – If there is no improvement in the monitored value during the specified number of epochs, the training stops.

• mode –

  One of {'auto', 'min', 'max'} is selected.

  • min mode ends the training when the decrease in the monitored value stops.

  • max mode, the training is terminated when the monitored values stop increasing.

  • auto mode, automatically estimated from the monitored values.

Methods

check_metrics(logs)
on_epoch_end(epoch[, logs])	Called when the epoch ends.

on_epoch_end(epoch, logs: Optional[Dict] = None) → None

Called when the epoch ends.

EarlyStoppingForTSUS

class enchanter.callbacks.EarlyStoppingForTSUS(data: torch.Tensor, target: torch.Tensor, classifier: sklearn.base.BaseEstimator = SVC(), monitor: str = 'accuracy', min_delta: float = 0.0, patience: int = 0, kfold: Optional[sklearn.model_selection._split.BaseCrossValidator] = None, mode: str = 'auto', grid_search: Optional[Dict[str, Any]] = None)

Bases: enchanter.callbacks.base.Callback

Early Stopping for Time Series Unsupervised Runner.

Examples

>>> import torch
>>> from enchanter.tasks import TimeSeriesUnsupervisedRunner
>>> x_train = torch.randn(32, 3, 128)
>>> y_train = torch.randint(0, high=4, size=(128, ))
>>> runner = TimeSeriesUnsupervisedRunner(
>>>     callbacks=[EarlyStoppingForTSUS(x_train, y_train)]
>>> )

Initializer

Parameters

• data – Data for training a classifier to evaluate the quality of the encoder’s output representation.

• target – Targets for training a classifier to evaluate the quality of the encoder’s output representation.

• classifier – A classifier for evaluating the quality of the output representation of the encoder.

• monitor –

• min_delta –

• patience –

• kfold –

• mode –

Methods

cross_val()
encode()
on_epoch_end(epoch[, logs])	Called when the epoch ends.

on_epoch_end(epoch, logs: Optional[Dict] = None)

Called when the epoch ends.

---

Logging

Provides an alternative logging method to comet_ml.Experiment.

BaseLogger

class enchanter.callbacks.BaseLogger

Bases: abc.ABC

Provides minimal compatibility with the comet_ml.Experiment, which is required to run Runner.

abstract end()

Use to indicate that the experiment is complete.

Warning

If you create your own Logger, you will need to implement this method.

abstract log_metric(name: str, value: Any, step: Optional[int] = None, epoch: Optional[int] = None, include_context: bool = True) → None

Logs a metric.

Parameters

• name – name of metric

• value –

• step –

• epoch –

• include_context –

Returns

None

Warning

If you create your own Logger, you will need to implement this method.

abstract log_metrics(dic: Dict, prefix: Optional[str] = None, step: Optional[int] = None, epoch: Optional[int] = None) → None

Logs a key, value dictionary of metrics.

See also

log_metric

Warning

If you create your own Logger, you will need to implement this method.

abstract log_parameter(name: str, value: Any, step: Optional[int] = None) → None

Logs a single hyper-parameter.

Parameters

• name – name of hyper-parameter

• value – value

• step –

Warning

If you create your own Logger, you will need to implement this method.

abstract log_parameters(dic: Dict, prefix: Optional[str] = None, step: Optional[int] = None) → None

Logs a key, value dictionary of hyper-parameters.

See also

log_peramter

Warning

If you create your own Logger, you will need to implement this method.

TensorBoardLogger

class enchanter.callbacks.TensorBoardLogger(*args, **kwargs)

Bases: enchanter.callbacks.loggers.BaseLogger

TenorBoardLogger is a module that supports the minimum logging in the environment where comet.ml cannot be used.

Examples

>>> from enchanter.tasks import ClassificationRunner
>>> model, optimizer, criterion = ...
>>> runner = ClassificationRunner(
>>>     model, optimizer, criterion, experiment=TensorBoardLogger()
>>> )

See the initializer argument for torch.utils.tensorboard.writer.SummaryWriter .

end()

Use to indicate that the experiment is complete.

log_metric(name: str, value: Any, step: Optional[int] = None, epoch: Optional[int] = None, include_context: bool = True) → None

Logs a metric.

Parameters

• name – name of metric

• value – value

• step – Used as the X axis when plotting on TensorBoard.

• epoch – Used as the X axis when plotting on TensorBoard.

• include_context –

Returns

None

log_metrics(dic: Dict, prefix: Optional[str] = None, step: Optional[int] = None, epoch: Optional[int] = None) → None

Logs a key, value dictionary of metrics.

See also

log_metric

log_parameter(name: Optional[str], value: Any, step: Optional[int] = None) → None

Logs a single hyper-parameter.

Parameters

• name – name of hyper-parameter

• value – value

• step – used as the X-axis when plotting on TensorBoard.

Returns:

log_parameters(dic: Dict, prefix: Optional[str] = None, step: Optional[int] = None) → Any

Logs a key, value dictionary of hyper-parameters.

See also

log_peramter