composer.callbacks.callback_hparams

Hyperparameters for callbacks.

Hparams

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

CallbackHparams	Base class for Callback hyperparameters.
CheckpointSaverHparams	CheckpointSaver hyperparameters.
GradMonitorHparams	GradMonitor hyperparamters.
LRMonitorHparams	LRMonitor hyperparameters.
MemoryMonitorHparams	MemoryMonitor hyperparameters.
SpeedMonitorHparams	SpeedMonitor hyperparameters.

class composer.callbacks.callback_hparams.CallbackHparams

Bases: yahp.hparams.Hparams, abc.ABC

Base class for Callback hyperparameters.

abstract initialize_object()

Initialize the callback.

Returns

Callback – An instance of the callback.

class composer.callbacks.callback_hparams.CheckpointSaverHparams(save_folder='{run_name}/checkpoints', filename='ep{epoch}-ba{batch}-rank{rank}', artifact_name='{run_name}/checkpoints/ep{epoch}-ba{batch}-rank{rank}', latest_filename='latest-rank{rank}', overwrite=False, weights_only=False, save_interval='1ep', num_checkpoints_to_keep=- 1)

Bases: composer.callbacks.callback_hparams.CallbackHparams

CheckpointSaver hyperparameters.

Parameters

• save_folder (str, optional) – See CheckpointSaver.

• filename (str, optional) – See CheckpointSaver.

• artifact_name (str, optional) – See CheckpointSaver.

• latest_filename (str, optional) – See CheckpointSaver.

• overwrite (str, optional) – See CheckpointSaver.

• weights_only (bool, optional) – See CheckpointSaver.

• num_checkpoints_to_keep (int, optional) – See CheckpointSaver.

• save_interval (str, optional) –

  Either a time-string or a path to a function.

  If a time-string, checkpoints will be saved according to this interval.

  If a path to a function, it should be of the format 'path.to.function:function_name'. The function should take (State, Event) and return a boolean indicating whether a checkpoint should be saved given the current state and event. The event will be either BATCH_CHECKPOINT or EPOCH_CHECKPOINT.

class composer.callbacks.callback_hparams.GradMonitorHparams(log_layer_grad_norms=False)

Bases: composer.callbacks.callback_hparams.CallbackHparams

GradMonitor hyperparamters.

Parameters

log_layer_grad_norms (bool, optional) – See GradMonitor for documentation.

initialize_object()

Initialize the GradMonitor callback.

Returns

GradMonitor – An instance of GradMonitor.

class composer.callbacks.callback_hparams.LRMonitorHparams

Bases: composer.callbacks.callback_hparams.CallbackHparams

LRMonitor hyperparameters.

There are no parameters as LRMonitor does not take any parameters.

initialize_object()

Initialize the LRMonitor callback.

Returns

LRMonitor – An instance of LRMonitor.

class composer.callbacks.callback_hparams.MLPerfCallbackHparams(root_folder, index, benchmark='resnet', target=0.759, division='open', metric_name='Accuracy', metric_label='eval', submitter='MosaicML', system_name=None, status='onprem', cache_clear_cmd=None)

Bases: composer.callbacks.callback_hparams.CallbackHparams

MLPerfCallback hyperparameters.

Parameters

• root_folder (str) – The root submission folder

• index (int) – The repetition index of this run. The filename created will be result_[index].txt.

• benchmark (str, optional) – Benchmark name. Currently only resnet supported.

• target (float, optional) – The target metric before the mllogger marks the stop of the timing run. Default: 0.759 (resnet benchmark).

• division (str, optional) – Division of submission. Currently only open division supported.

• metric_name (str, optional) – name of the metric to compare against the target. Default: Accuracy.

• metric_label (str, optional) – label name. The metric will be accessed via state.current_metrics[metric_label][metric_name].

• submitter (str, optional) – Submitting organization. Default: MosaicML.

• system_name (str, optional) – Name of the system (e.g. 8xA100_composer). If not provided, system name will default to [world_size]x[device_name]_composer, e.g. 8xNVIDIA_A100_80GB_composer.

• status (str, optional) – Submission status. One of (onprem, cloud, or preview). Default: "onprem".

• cache_clear_cmd (str, optional) – Command to invoke during the cache clear. This callback will call subprocess(cache_clear_cmd). Default is disabled (None)

initialize_object()

Initialize the MLPerf Callback.

Returns

MLPerfCallback – An instance of MLPerfCallback

class composer.callbacks.callback_hparams.MemoryMonitorHparams

Bases: composer.callbacks.callback_hparams.CallbackHparams

MemoryMonitor hyperparameters.

There are no parameters as MemoryMonitor does not take any parameters.

initialize_object()

Initialize the MemoryMonitor callback.

Returns

MemoryMonitor – An instance of MemoryMonitor.

class composer.callbacks.callback_hparams.SpeedMonitorHparams(window_size=100)

Bases: composer.callbacks.callback_hparams.CallbackHparams

SpeedMonitor hyperparameters.

Parameters

window_size (int, optional) – See SpeedMonitor for documentation.

initialize_object()

Initialize the SpeedMonitor callback.

Returns

SpeedMonitor – An instance of SpeedMonitor.