tf.VariableAggregation

TensorFlow 2 version View source on GitHub

Indicates how a distributed variable will be aggregated.

tf.distribute.Strategy distributes a model by making multiple copies (called "replicas") acting data-parallel on different elements of the input batch. When performing some variable-update operation, say var.assign_add(x), in a model, we need to resolve how to combine the different values for x computed in the different replicas.

  • NONE: This is the default, giving an error if you use a variable-update operation with multiple replicas.
  • SUM: Add the updates across replicas.
  • MEAN: Take the arithmetic mean ("average") of the updates across replicas.
  • ONLY_FIRST_REPLICA: This is for when every replica is performing the same update, but we only want to perform the update once. Used, e.g., for the global step counter.
  • ONLY_FIRST_TOWER: Deprecated alias for ONLY_FIRST_REPLICA.

Class Variables

  • MEAN
  • NONE
  • ONLY_FIRST_REPLICA
  • SUM