ml.utils.exceptions

Custom exception types.

exception ml.utils.exceptions.NaNError[source]

Bases: Exception

Raised when NaNs are detected in the model parameters.

exception ml.utils.exceptions.EpochDoneError[source]

Bases: Exception

Raised when an epoch is done.

exception ml.utils.exceptions.TrainingFinishedError[source]

Bases: Exception

Raised when training is finished.

exception ml.utils.exceptions.MinGradScaleError[source]

Raised when the minimum gradient scale is reached.

This is a subclass of TrainingFinishedError because it indicates that training is finished and causes the post-training hooks to be run.