tf.distribute.experimental.PreemptionCheckpointHandler  |  TensorFlow v2.16.1 (original) (raw)

Preemption and error handler for synchronous training.

tf.distribute.experimental.PreemptionCheckpointHandler(
    cluster_resolver,
    checkpoint_or_checkpoint_manager,
    checkpoint_dir=None,
    termination_config=None
)

A PreemptionCheckpointHandler coordinates all workers to save a checkpoint upon receiving a preemption signal. It also helps disseminate application error messages accurately among the cluster. When aPreemptionCheckpointHandler object is created, it restores values from the latest checkpoint file if any exists.

Right after the initialization, the object starts to watch out for termination signal for any member in the cluster. If receiving a signal, the next time the worker executes PreemptionCheckpointHandler.run, thePreemptionCheckpointHandler will align all workers to save a checkpoint. Then, if an exit_fn is configured viatf.distribute.experimental.TerminationConfig, it will be invoked. Otherwise, the process will simply exit and later the platform should restart it.

For users of tf.distribute.MultiWorkerMirroredStrategy, the core API isPreemptionCheckpointHandler.run:

strategy = tf.distribute.MultiWorkerMirroredStrategy()

trained_epoch = tf.Variable(initial_value=tf.constant(0, dtype=tf.dtypes.int64), name='epoch')
step_in_epoch = tf.Variable(initial_value=tf.constant(0, dtype=tf.dtypes.int64), name='step_in_epoch')

with strategy.scope():
  dataset, model, optimizer = ...

  checkpoint = tf.train.Checkpoint(optimizer=optimizer,
                                   model=model,
                                   trained_epoch=trained_epoch,
                                   step_in_epoch=step_in_epoch)

  preemption_checkpoint_handler = tf.distribute.experimental.PreemptionCheckpointHandler(cluster_resolver, checkpoint, checkpoint_dir)

while trained_epoch.numpy() < NUM_EPOCH:

  while step_in_epoch.numpy() < STEPS_PER_EPOCH:

    # distributed_train_function contains a call to strategy.run.
    loss += preemption_checkpoint_handler.run(distributed_train_function, args=(next(iterator),))
    # For users of MultiWorkerMirroredStrategy, usually
    # STEPS_PER_TRAIN_FUNCTION = 1.
    step_in_epoch.assign_add(STEPS_PER_TRAIN_FUNCTION)
    ...

  epoch.assign_add(1)
  step_in_epoch.assign(0)

For users of tf.distribute.TPUStrategy, the core APIs arePreemptionCheckpointHandler.run andPreemptionCheckpointHandler.watch_preemption_scope:

strategy = tf.distribute.TPUStrategy(tpu_cluster_resolver)

# Rest of TPU init omitted, see documentation for TPUSTrategy.

with preemption_checkpoint_handler.watch_preemption_scope():
  while trained_epoch.numpy() < NUM_EPOCH:

    while step_in_epoch.numpy() < STEPS_PER_EPOCH:

      # distributed_train_function contains a call to strategy.run.
      loss += preemption_checkpoint_handler.run(distributed_train_function, args=(next(iterator),))

      # For users of TPUStrategy, usually STEPS_PER_TRAIN_FUNCTION >> 1 since
      # clustering multiple steps within a tf.function amortizes the overhead
      # of launching a multi-device function on TPU Pod.
      step_in_epoch.assign_add(STEPS_PER_TRAIN_FUNCTION)
      ...

    epoch.assign_add(1)
    step_in_epoch.assign(0)

Not all interruptions come with advance notice so that thePreemptionCheckpointHandler can handle them, e.g., those caused by hardware failure. For a user who saves checkpoints for these cases themselves outside the PreemptionCheckpointHandler, if they are using atf.train.CheckpointManager, pass it as thecheckpoint_or_checkpoint_manager argument to thePreemptionCheckpointHandler. If they do not have atf.train.CheckpointManager but are directly working withtf.train.Checkpoint, we advise saving the checkpoints in the directory that's passed as the checkpoint_dir argument. In this way, at the program beginning, PreemptionCheckpointHandler can restore the latest checkpoint from the directory, no matter it's saved by the user themselves or saved by the PreemptionCheckpointHandler before preemption happens.

A note on the platform:

PreemptionCheckpointHandler can only handle the kind of termination with advance notice. For now, the API recognizes the termination signal for CPU, GPU, and TPU on Google Borg and CPU and GPU on the Google Cloud Platform. In these cases, PreemptionCheckpointHandler will automatically adopt the correct preemption/maintenance notification detection mechanism. Users of other platforms can configure a detection monitoring behavior through thetf.distribute.experimental.TerminationConfig. Customization for the exit behavior and grace period length could also be done here.

Methods

run

View source

run(
    distributed_train_function, *args, **kwargs
)

Runs a training function with error and preemption handling.

This function handles the preemption signal from any peer in the cluster by saving the training progress and exiting gracefully. It will also broadcase any program error encountered during the execution ofdistributed_train_function to all workers so that they can raise the same error.

The distributed_train_function argument should be a distributed train function (i.e., containing a call to tf.distribute.Strategy.run). Fortf.distribute.MultiWorkerMirroredStrategy users, we recommend passing in a single-step distributed_train_function toPreemptionCheckpointHandler.run so that the checkpoint can be saved in time in case a preemption signal or maintenance notice is sent.

Besides the preemption and error handling part,PreemptionCheckpointHandler.run(distributed_train_function, *args, **kwargs) has the same effect and output asdistributed_train_function(*args, **kwargs). distributed_train_functioncan return either some or no result. The following is a shortened example:

@tf.function
def distributed_train_step(iterator):
  # A distributed single-step training function.

  def step_fn(inputs):
    # A per-replica single-step training function.
    x, y = inputs
    ...
    return loss

  per_replica_losses = strategy.run(step_fn, args=(next(iterator),))
  return strategy.reduce(
      tf.distribute.ReduceOp.SUM, per_replica_losses, axis=None)

for epoch in range(preemption_handler.total_run_calls // STEPS_PER_EPOCH,
                   EPOCHS_TO_RUN):
  iterator = iter(multi_worker_dataset)
  total_loss = 0.0
  num_batches = 0

  for step in range(preemption_handler.total_run_calls % STEPS_PER_EPOCH,
                    STEPS_PER_EPOCH):
    total_loss += preemption_handler.run(distributed_train_step)
    num_batches += 1

  train_loss = total_loss / num_batches
  print('Epoch: %d, train_loss: %f.' %(epoch.numpy(), train_loss))

  train_accuracy.reset_states()

save_checkpoint_if_preempted

View source

save_checkpoint_if_preempted(
    *args, **kwargs
)

Saves a checkpoint if a preemption signal has been made available.

This is an alternative API for PreemptionCheckpointHandler.run andPreemptionCheckpointHandler.watch_preemption_scope. This method works for both tf.distribute.MultiWorkerMirroredStrategy andtf.distribute.TPUStrategy. However, for TPUStrategy, this method will add a synchronization point between workers and the coordinator and thus may have performance implication. If this is a concern, use the combination of PreemptionCheckpointHandler.watch_preemption_scope andPreemptionCheckpointHandler.run instead.

strategy = tf.distribute.TPUStrategy(tpu_cluster_resolver)
# initialization omitted

with strategy.scope():
  # Save in the checkpoint.
  trained_step = tf.Variable(initial_value=tf.constant(0, dtype=tf.dtypes.int64), name='trained_step', aggregation=tf.VariableAggregation.ONLY_FIRST_REPLICA)

  checkpoint_manager = tf.train.CheckpointManager(checkpoint, directory, max_to_keep=1)
  preemption_handler = tf.distribute.experimental.PreemptionCheckpointHandler(cluster_resolver, checkpoint_manager)

while trained_step.numpy() < NUM_STEPS:
  # Train STEPS_IN_FUNCTION steps at once.
  train_multi_step_function()
  trained_step.assign_add(STEPS_IN_FUNCTION)
  preemption_handler.save_checkpoint_if_preempted()

watch_preemption_scope

View source

@tf_contextlib.contextmanager watch_preemption_scope()

Syncs error and maybe save checkpoint for usage with TPUStrategy.

Example usage:

with preemption_checkpoint_handler.watch_preemption_scope():
  while trained_step.numpy() < NUM_STEPS:

    # distributed_train_function contains a call to strategy.run.
    loss += preemption_checkpoint_handler.run(distributed_train_function, args=(next(iterator),))
    trained_step.assign_add(STEPS_PER_TRAIN_FUNCTION)

In this workflow, PreemptionCheckpointHandler.run will flag preemption signal received, and watch_preemption_scope will handle the preemption signal by saving a checkpoint and then either exit to restart or execute a user-passed exit_fn in tf.distribute.experimental.TerminationConfig. If no preemption signal is received during execution of ops and function inside the scope, watch_preemption_scope ensures the completion of all async op and function execution when exiting and will raises exceptions if async execution results in an error state.