drytorch.core.exceptions

Module containing internal exceptions for the drytorch package.

Exceptions

`AccessOutsideScopeError`(*args)	Raised when an operation is attempted outside an experiment scope.
`CannotStoreOutputWarning`(error)	Warning raised when output cannot be stored due to an error.
`CheckpointNotInitializedError`(*args)	Raised when attempting to use a checkpoint without a registered model.
`ComputedBeforeUpdatedWarning`(calculator)	Warning raised when compute method is called before updating.
`ComputedMetricsTypeError`(computed_metrics_type)	Raised when computed metrics have an unexpected type.
`ConvergenceError`(criterion)	Raised when a module fails to converge during training.
`DatasetHasNoLengthError`(*args)	Raised when a dataset does not implement the __len__ method.
`DeviceMismatchError`(metric_name, ...)	Raised when the metrics device does not match the expected device.
`DistributedDatasetNotDivisibleWarning`(name, ...)	Warning raised when the dataset cannot be equally distributed.
`DistributedStorageWarning`(error)	Warning raised when the distributed storage is not synchronized.
`DryTorchError`(*args)	Base exception class for all drytorch package exceptions.
`DryTorchWarning`(*args)	Base warning class for all drytorch package warnings.
`EpochNotFoundError`(epoch, checkpoint_directory)	Raised when no saved model is found in the checkpoint directory.
`ExperimentalFeatureWarning`(feature)	Warning raised when an experimental feature is used.
`FailedOptionalImportWarning`(package_name)	Warning raised when an optional dependency fails to import.
`FuncNotApplicableError`(func_name, type_name)	Raised when a function cannot be applied to a specific type.
`LossNotScalarError`(size)	Raised when a loss value is not a scalar tensor.
`MetricNotFoundError`(source_name, metric_name)	Raised when a requested metric is not found in the specified source.
`MissingParamError`(module_names, lr_param_groups)	Raised when parameter groups are missing required parameters.
`ModelDeviceMismatchError`(*args)	Raised when the metrics device does not match the model device.
`ModelNotFoundError`(checkpoint_directory)	Raised when no saved model is found in the checkpoint directory.
`ModuleAlreadyRegisteredError`(model_name, ...)	Raised when trying to access a model that has already been registered.
`ModuleNotDistributedWarning`(*args)	Warning raised when a model is not distributed.
`ModuleNotRegisteredError`(model_name, ...)	Raised an actor tries to access a module that hasn't been registered.
`NameAlreadyRegisteredError`(name)	Raised when attempting to register a name already in use.
`NamedTupleOnlyError`(tuple_type)	Raised when operations require a named tuple and not a subclass.
`NestedScopeError`(current_exp_name, new_exp_name)	Raised when attempting to nest an experiment scope within another one.
`NoActiveExperimentError`([experiment_name, ...])	Raised when no experiment is currently active.
`NoPreviousRunsWarning`(*args)	Attempted to resume the last run, but none were found.
`NotExistingRunWarning`(run_id)	Attempted to resume a not existing run.
`ObjectiveSyncWarning`(issue, recommend)	Warning for metric synchronization configuration issues.
`OptimizerNotLoadedWarning`(error)	Warning raised when the optimizer has not been correctly loaded.
`PastEpochWarning`(selected_epoch, current_epoch)	Warning raised when training is requested for a past epoch.
`RecursionWarning`(*args)	Warning raised when recursive objects obstruct metadata extraction.
`ResultNotAvailableError`(*args)	Raised when trying to access a result before the hook has been called.
`RunAlreadyCompletedWarning`(*args)	Warning raised when a run is stopped after completion.
`RunAlreadyRecordedError`(run_id, exp_name)	Error raised when attempting to record a run multiple times.
`RunAlreadyRunningWarning`(*args)	Warning raised when a run is started when already running.
`RunNotRecordedError`(run_id)	Raised when attempting to update a run that is not registered.
`RunNotStartedWarning`(*args)	Warning raised when a run is stopped before being started.
`TerminatedTrainingWarning`(*args)	Warning raised when training is attempted after termination.
`TrackerAlreadyRegisteredError`(tracker_name, ...)	Raised when attempting to register an already registered tracker.
`TrackerError`(tracker, tracker_msg)	Exception raised by tracker objects during experiment tracking.
`TrackerExceptionWarning`(subscriber_name, error)	Warning raised when a tracker encounters an error and is skipped.
`TrackerNotUsedError`(tracker_name)	Raised when trying to access a tracker that is not registered.

exception AccessOutsideScopeError(*args: Any)[source]

Bases: DryTorchError

Raised when an operation is attempted outside an experiment scope.

Initialize.

Parameters:: *args (Any) – arguments to be formatted into the message template.
Return type:: None

exception CannotStoreOutputWarning(error: BaseException)[source]

Bases: DryTorchWarning

Warning raised when output cannot be stored due to an error.

Initialize.

Parameters:: error (BaseException) – the error that prevented output storage.
Return type:: None

exception CheckpointNotInitializedError(*args: Any)[source]

Bases: DryTorchError

Raised when attempting to use a checkpoint without a registered model.

Initialize.

Parameters:: *args (Any) – arguments to be formatted into the message template.
Return type:: None

exception ComputedBeforeUpdatedWarning(calculator: Any)[source]

Bases: DryTorchWarning

Warning raised when compute method is called before updating.

Initialize.

Parameters:: calculator (Any) – the calculator object that was computed before updating.
Return type:: None

exception ComputedMetricsTypeError(computed_metrics_type: type)[source]

Bases: DryTorchError

Raised when computed metrics have an unexpected type.

Initialize.

Parameters:: computed_metrics_type (type) – the actual type of the computed metrics.
Return type:: None

exception ConvergenceError(criterion: float)[source]

Bases: DryTorchError

Raised when a module fails to converge during training.

Initialize.

Parameters:: criterion (float) – the convergence criterion that was not met.
Return type:: None

exception DatasetHasNoLengthError(*args: Any)[source]

Bases: DryTorchError

Raised when a dataset does not implement the __len__ method.

Initialize.

Parameters:: *args (Any) – arguments to be formatted into the message template.
Return type:: None

exception DeviceMismatchError(metric_name: str, metric_device: device, target_device: device)[source]

Bases: DryTorchError

Raised when the metrics device does not match the expected device.

Initialize.

Parameters:

metric_name (str) – the name of the metric.
metric_device (device) – the device of the output tensor.
target_device (device) – the device of the model.

Return type:

None

exception DistributedDatasetNotDivisibleWarning(name: str, len_dataset: int, n_processes: int)[source]

Bases: DryTorchWarning

Warning raised when the dataset cannot be equally distributed.

Initialize.

Parameters:

name (str) – the name of the actor experiencing the issue.
len_dataset (int) – the size of the dataset.
n_processes (int) – the number of processes used in distributed processing.

Return type:

None

exception DistributedStorageWarning(error: BaseException)[source]

Bases: DryTorchWarning

Warning raised when the distributed storage is not synchronized.

Initialize.

Parameters:: error (BaseException) – the error that occurred while synchronizing the storage.
Return type:: None

exception DryTorchError(*args: Any)[source]

Bases: Exception

Base exception class for all drytorch package exceptions.

Initialize.

Parameters:: *args (Any) – arguments to be formatted into the message template.
Return type:: None

exception DryTorchWarning(*args: Any)[source]

Bases: UserWarning

Base warning class for all drytorch package warnings.

Initialize.

Parameters:: *args (Any) – arguments to be formatted into the message template.
Return type:: None

exception EpochNotFoundError(epoch: int, checkpoint_directory: Path)[source]

Bases: DryTorchError

Raised when no saved model is found in the checkpoint directory.

Initialize.

Parameters:

epoch (int) – the epoch that was not found.
checkpoint_directory (Path) – the directory path where no model was found.

Return type:

None

exception ExperimentalFeatureWarning(feature: str)[source]

Bases: DryTorchWarning

Warning raised when an experimental feature is used.

Initialize.

Parameters:: feature (str) – the experimental feature that was used.
Return type:: None

exception FailedOptionalImportWarning(package_name: str)[source]

Bases: DryTorchWarning

Warning raised when an optional dependency fails to import.

Initialize.

Parameters:: package_name (str) – the name of the package that failed to import.
Return type:: None

exception FuncNotApplicableError(func_name: str, type_name: str)[source]

Bases: DryTorchError

Raised when a function cannot be applied to a specific type.

Initialize.

Parameters:

func_name (str) – the name of the function that cannot be applied.
type_name (str) – the name of the type that doesn’t support the function.

Return type:

None

exception LossNotScalarError(size: Size)[source]

Bases: DryTorchError

Raised when a loss value is not a scalar tensor.

Initialize.

Parameters:: size (Size) – the actual size of the non-scalar loss tensor.
Return type:: None

exception MetricNotFoundError(source_name: str, metric_name: str)[source]

Bases: DryTorchError

Raised when a requested metric is not found in the specified source.

Initialize.

Parameters:

source_name (str) – the name of the source where the metric was not found.
metric_name (str) – the name of the metric that was not found.

Return type:

None

exception MissingParamError(module_names: list[str], lr_param_groups: list[str])[source]

Bases: DryTorchError

Raised when parameter groups are missing required parameters.

Initialize.

Parameters:

module_names (list[str]) – list of module names that should have parameters.
lr_param_groups (list[str]) – group names in the parameter learning rate config.

Return type:

None

exception ModelDeviceMismatchError(*args: Any)[source]

Bases: DryTorchError

Raised when the metrics device does not match the model device.

Initialize.

Parameters:: *args (Any) – arguments to be formatted into the message template.
Return type:: None

exception ModelNotFoundError(checkpoint_directory: Path)[source]

Bases: DryTorchError

Raised when no saved model is found in the checkpoint directory.

Initialize.

Parameters:: checkpoint_directory (Path) – the directory path where no model was found.
Return type:: None

exception ModuleAlreadyRegisteredError(model_name: str, exp_name: str, run_id: str)[source]

Bases: DryTorchError

Raised when trying to access a model that has already been registered.

Initialize.

Parameters:

model_name (str) – the name of the model that was not registered.
exp_name (str) – the name of the current experiment.
run_id (str) – the current run’s id.

Return type:

None

exception ModuleNotDistributedWarning(*args: Any)[source]

Bases: DryTorchWarning

Warning raised when a model is not distributed.

Initialize.

Parameters:: *args (Any) – arguments to be formatted into the message template.
Return type:: None

exception ModuleNotRegisteredError(model_name: str, exp_name: str, run_id: str)[source]

Bases: DryTorchError

Raised an actor tries to access a module that hasn’t been registered.

Initialize.

Parameters:

model_name (str) – the name of the model that was not registered.
exp_name (str) – the name of the current experiment.
run_id (str) – the current run’s id.

Return type:

None

exception NameAlreadyRegisteredError(name: str)[source]

Bases: DryTorchError

Raised when attempting to register a name already in use.

Initialize.

Parameters:: name (str) – the name that is already registered.
Return type:: None

exception NamedTupleOnlyError(tuple_type: str)[source]

Bases: DryTorchError

Raised when operations require a named tuple and not a subclass.

Initialize.

Parameters:: tuple_type (str) – the actual type of the tuple that was provided.
Return type:: None

exception NestedScopeError(current_exp_name: str, new_exp_name: str)[source]

Bases: DryTorchError

Raised when attempting to nest an experiment scope within another one.

Initialize.

Parameters:

current_exp_name (str) – the name of the currently active experiment.
new_exp_name (str) – the name of the experiment that cannot be started.

Return type:

None

exception NoActiveExperimentError(experiment_name: str | None = None, experiment_class: type | None = None)[source]

Bases: DryTorchError

Raised when no experiment is currently active.

Initialize.

Parameters:

experiment_name (str | None) – specifies experiment’s name.
experiment_class (type | None) – specifies experiment’s name.

Return type:

None

exception NoPreviousRunsWarning(*args: Any)[source]

Bases: DryTorchWarning

Attempted to resume the last run, but none were found.

Initialize.

Parameters:: *args (Any) – arguments to be formatted into the message template.
Return type:: None

exception NotExistingRunWarning(run_id: str)[source]

Bases: DryTorchWarning

Attempted to resume a not existing run.

Initialize.

Parameters:: run_id (str) – the id of the run that was not found.
Return type:: None

exception ObjectiveSyncWarning(issue: str, recommend: str)[source]

Bases: DryTorchWarning

Warning for metric synchronization configuration issues.

Initialize.

Parameters:

issue (str) – the issue that was encountered with the objective.
recommend (str) – the recommended action to fix the issue.

Return type:

None

exception OptimizerNotLoadedWarning(error: BaseException)[source]

Bases: DryTorchWarning

Warning raised when the optimizer has not been correctly loaded.

Initialize.

Parameters:: error (BaseException) – the error that occurred while loading the optimizer.
Return type:: None

exception PastEpochWarning(selected_epoch: int, current_epoch: int)[source]

Bases: DryTorchWarning

Warning raised when training is requested for a past epoch.

Initialize.

Parameters:

selected_epoch (int) – the epoch that training was requested until.
current_epoch (int) – the current epoch number.

Return type:

None

exception RecursionWarning(*args: Any)[source]

Bases: DryTorchWarning

Warning raised when recursive objects obstruct metadata extraction.

Initialize.

Parameters:: *args (Any) – arguments to be formatted into the message template.
Return type:: None

exception ResultNotAvailableError(*args: Any)[source]

Bases: DryTorchError

Raised when trying to access a result before the hook has been called.

Initialize.

Parameters:: *args (Any) – arguments to be formatted into the message template.
Return type:: None

exception RunAlreadyCompletedWarning(*args: Any)[source]

Bases: DryTorchWarning

Warning raised when a run is stopped after completion.

Initialize.

Parameters:: *args (Any) – arguments to be formatted into the message template.
Return type:: None

exception RunAlreadyRecordedError(run_id: str, exp_name: str)[source]

Bases: DryTorchError

Error raised when attempting to record a run multiple times.

Initialize.

Parameters:

run_id (str) – the id of the run that is already recorded.
exp_name (str) – the name of the experiment where to record the run.

Return type:

None

exception RunAlreadyRunningWarning(*args: Any)[source]

Bases: DryTorchWarning

Warning raised when a run is started when already running.

Initialize.

Parameters:: *args (Any) – arguments to be formatted into the message template.
Return type:: None

exception RunNotRecordedError(run_id: str)[source]

Bases: DryTorchError

Raised when attempting to update a run that is not registered.

Constructor.

Parameters:: run_id (str) – the id of the run that is not registered.
Return type:: None

exception RunNotStartedWarning(*args: Any)[source]

Bases: DryTorchWarning

Warning raised when a run is stopped before being started.

Initialize.

Parameters:: *args (Any) – arguments to be formatted into the message template.
Return type:: None

exception TerminatedTrainingWarning(*args: Any)[source]

Bases: DryTorchWarning

Warning raised when training is attempted after termination.

Initialize.

Parameters:: *args (Any) – arguments to be formatted into the message template.
Return type:: None

exception TrackerAlreadyRegisteredError(tracker_name: str, exp_name: str)[source]

Bases: DryTorchError

Raised when attempting to register an already registered tracker.

Initialize.

Parameters:

tracker_name (str) – the name of the tracker that is already registered.
exp_name (str) – the name of the experiment where to register the tracker.

Return type:

None

exception TrackerError(tracker: Any, tracker_msg: str)[source]

Bases: DryTorchError

Exception raised by tracker objects during experiment tracking.

Initialize.

Parameters:

tracker (Any) – the tracker object that encountered the error.
tracker_msg (str) – the error message from the tracker.

Return type:

None

exception TrackerExceptionWarning(subscriber_name: str, error: BaseException)[source]

Bases: DryTorchWarning

Warning raised when a tracker encounters an error and is skipped.

Constructor.

Parameters:

subscriber_name (str) – the name of the tracker that encountered the error.
error (BaseException) – the error that occurred in the tracker.

Return type:

None

exception TrackerNotUsedError(tracker_name: str)[source]

Bases: DryTorchError

Raised when trying to access a tracker that is not registered.

Initialize.

Parameters:: tracker_name (str) – the name of the tracker that is not registered.
Return type:: None