composer.callbacks

Modules

composer.callbacks.callback_hparams	Hyperparameters for callbacks.
composer.callbacks.checkpoint_saver	Callback to save checkpoints during training.
composer.callbacks.early_stopper	Early stopping callback.
composer.callbacks.grad_monitor	Monitor gradient during training.
composer.callbacks.lr_monitor	Monitor learning rate during training.
composer.callbacks.memory_monitor	Log memory usage during training.
composer.callbacks.mlperf	composer.callbacks.mlperf
composer.callbacks.speed_monitor	Monitor throughput during training.
composer.callbacks.threshold_stopper	Threshold stopping callback.

Callbacks that run at each training loop Event.

Each callback inherits from the Callback base class. See detailed description and examples for writing your own callbacks at the Callback base class.

Classes

CheckpointSaver	Callback to save checkpoints.
EarlyStopper	This callback tracks a training or evaluation metric and halts training if the metric does not improve within a given interval.
GradMonitor	Computes and logs the L2 norm of gradients on the AFTER_TRAIN_BATCH event.
LRMonitor	Logs the learning rate.
MLPerfCallback	Creates a compliant results file for MLPerf Training benchmark.
MemoryMonitor	Logs the memory usage of the model.
SpeedMonitor	Logs the training throughput.
ThresholdStopper	This callback tracks a training or evaluation metric and halts training when the metric value reaches a certain threshold.

Hparams

These classes are used with yahp for YAML-based configuration.

CallbackHparams	Base class for Callback hyperparameters.
CheckpointSaverHparams	CheckpointSaver hyperparameters.
EarlyStopperHparams	EarlyStopper hyperparameters.
GradMonitorHparams	GradMonitor hyperparamters.
LRMonitorHparams	LRMonitor hyperparameters.
MLPerfCallbackHparams	MLPerfCallback hyperparameters.
MemoryMonitorHparams	MemoryMonitor hyperparameters.
SpeedMonitorHparams	SpeedMonitor hyperparameters.
ThresholdStopperHparams	ThresholdStopper hyperparameters.