allennlp.training.callbacks.update_learning_rate#

UpdateLearningRate#

UpdateLearningRate(self, learning_rate_scheduler:allennlp.training.learning_rate_schedulers.learning_rate_scheduler.LearningRateScheduler) -> None

Callback that runs the learning rate scheduler.

Parameters

  • learning_rate_scheduler : LearningRateScheduler The scheduler to handler.

get_training_state#

UpdateLearningRate.get_training_state(self) -> dict

We need to persist the learning_rate_scheduler state as training state.