W3cubDocs

/TensorFlow 2.3

tf.distribute.experimental.ValueContext

A class wrapping information needed by a distribute function.

This is a context class that is passed to the value_fn in strategy.experimental_distribute_values_from_function and contains information about the compute replicas. The num_replicas_in_sync and replica_id can be used to customize the value on each replica.

Example usage:

  1. Directly constructed.
def value_fn(context):
  return context.replica_id_in_sync_group/context.num_replicas_in_sync
context = tf.distribute.experimental.ValueContext(
  replica_id_in_sync_group=2, num_replicas_in_sync=4)
per_replica_value = value_fn(context)
per_replica_value
0.5
  1. Passed in by experimental_distribute_values_from_function.
strategy = tf.distribute.MirroredStrategy()
def value_fn(value_context):
  return value_context.num_replicas_in_sync
distributed_values = (
     strategy.experimental_distribute_values_from_function(
       value_fn))
local_result = strategy.experimental_local_results(distributed_values)
local_result
(1,)
Args
replica_id_in_sync_group the current replica_id, should be an int in [0,num_replicas_in_sync).
num_replicas_in_sync the number of replicas that are in sync.
Attributes
num_replicas_in_sync Returns the number of compute replicas in sync.
replica_id_in_sync_group Returns the replica ID.

© 2020 The TensorFlow Authors. All rights reserved.
Licensed under the Creative Commons Attribution License 3.0.
Code samples licensed under the Apache 2.0 License.
https://www.tensorflow.org/versions/r2.3/api_docs/python/tf/distribute/experimental/ValueContext