Save the date! Google I/O returns May 18-20 Register now


Synchronous training across multiple replicas on one machine.

Inherits From: Strategy

Used in the notebooks

Used in the guide Used in the tutorials

This strategy is typically used for training on one machine with multiple GPUs. For TPUs, use tf.distribute.TPUStrategy. To use MirroredStrategy with multiple workers, please refer to tf.distribute.experimental.MultiWorkerMirroredStrategy.

For example, a variable created under a MirroredStrategy is a MirroredVariable. If no devices are specified in the constructor argument of the strategy then it will use all the available GPUs. If no GPUs are found, it will use the available CPUs. Note that TensorFlow treats all CPUs on a machine as a single device, and uses threads internally for parallelism.

strategy = tf.distribute.MirroredStrategy(["GPU:0", "GPU:1"])
with strategy.scope():
  x = tf.Variable(1.)
  0: <tf.Variable ... shape=() dtype=float32, numpy=1.0>,
  1: <tf.Variable ... shape=() dtype=float32, numpy=1.0>

While using distribution strategies, all the variable creation should be done within the strategy's scope. This will replicate the variables across all the replicas and keep them in sync using an all-reduce algorithm.

Variables created inside a MirroredStrategy which is wrapped with a tf.function are still MirroredVariables.

x = []
@tf.function  # Wrap the function with tf.function.
def create_variable():
  if not x:
  return x[0]
strategy = tf.distribute.MirroredStrategy(["GPU:0", "GPU:1"])
with strategy.scope():
  _ = create_variable()
  0: <tf.Variable ... shape=() dtype=float32, numpy=1.0>,
  1: <tf.Variable ... shape=() dtype=float32, numpy=1.0>

experimental_distribute_dataset can be used to distribute the dataset across the replicas when writing your own training loop. If you are using .fit and .compile methods available in tf.keras, then tf.keras will handle the distribution for you.

For example:

my_strategy = tf.distribute.MirroredStrategy()
with my_strategy.scope():
  def distribute_train_epoch(dataset):
    def replica_fn(input):
      # process input and return result
      return result

    total_result = 0
    for x in dataset:
      per_replica_result =, args=(x,))
      total_result += my_strategy.reduce(tf.distribute.ReduceOp.SUM,
                                         per_replica_result, axis=None)
    return total_result

  dist_dataset = my_strategy.experimental_distribute_dataset(dataset)
  for _ in range(EPOCHS):
    train_result = distribute_train_epoch(dist_dataset)

devices a list of device strings such as ['/gpu:0', '/gpu:1']. If None, all available GPUs are used. If no GPUs are found, CPU is used.
cross_device_ops optional, a descedant of CrossDeviceOps. If this is not set, NcclAllReduce() will be used by default. One would customize this if NCCL isn't available or if a special implementation that exploits the particular hardware is available.

cluster_resolver Returns the cluster resolver associated with this strategy.

In general, when using a multi-worker tf.distribute strategy such as tf.distribute.experimental.MultiWorkerMirroredStrategy or tf.distribute.TPUStrategy(), there is a tf.distribute.cluster_resolver.ClusterResolver associated with the strategy used, and such an instance is returned by this property.

Strategies that intend to have an associated tf.distribute.cluster_resolver.ClusterResolver must set the relevant attribute, or override this property; otherwise, None is returned by default. Those strategies should also provide information regarding what is returned by this property.

Single-worker strategies usually do not have a tf.distribute.cluster_resolver.ClusterResolver, and in those cases this property will return None.

The tf.distribute.cluster_resolver.ClusterResolver may be useful when the user needs to access information such as the cluster spec, task type or task id. For example,

os.environ['TF_CONFIG'] = json.dumps({
'cluster': {
'worker': ["localhost:12345", "localhost:23456"],
'ps': ["localhost:34567"]
'task': {'type': 'worker', 'index': 0}

# This implicitly uses TF_CONFIG for the cluster and current task info.
strategy = tf.distribute.experimental.MultiWorkerMirroredStrategy()


if strategy.cluster_resolver.task_type == 'worker':
# Perform something that's only applicable on workers. Since we set this
# as a worker above, this block will run on this particular instance.
elif strategy.cluster_resolver.task_type == 'ps':
# Perform something that's only applicable on parameter servers. Since we
# set this as a worker above, this block will not run on this particular
# instance.

For more information, please see tf.distribute.cluster_resolver.ClusterResolver's API docstring.

extended tf.distribute.StrategyExtended with additional methods.
num_replicas_in_sync Returns number of replicas over which gradients are aggregated.



View source

Distributes instances created by calls to dataset_fn.

The argument dataset_fn that users pass in is an input function that has a tf.distribute.InputContext argument and returns a instance. It is expected that the returned dataset from dataset_fn is already batched by per-replica batch size (i.e. global batch size divided by the number of replicas in sync) and sharded. tf.distribute.Strategy.distribute_datasets_from_function does not batch or shard the instance returned from the input function. dataset_fn will be called on the CPU device of each of the workers and each generates a dataset where every replica on that worker will dequeue one batch of inputs (i.e. if a worker has two replicas, two batches will be dequeued from the Dataset every step).

This method can be used for several purposes. First, it allows you to specify your own batching and sharding logic. (In contrast, tf.distribute.experimental_distribute_dataset does batching and sharding for you.) For example, where experimental_distribute_dataset is unable to shard the input files, this method might be used to manually shard the dataset (avoiding the slow fallback behavior in experimental_distribute_dataset). In cases where the dataset is infinite, this sharding can be done by creating dataset replicas that differ only in their random seed.

The dataset_fn should take an tf.distribute.InputContext instance where information about batching and input replication can be accessed.

You can use element_spec property of the tf.distribute.DistributedDataset returned by this API to query the tf.TypeSpec of the elements returned by the iterator. This can be used to set the input_signature property of a tf.function. Follow tf.distribute.DistributedDataset.element_spec to see an example.

For a tutorial on more usage and properties of this method, refer to the tutorial on distributed input). If you are interested in last partial batch handling, read this section.

dataset_fn A function taking a tf.distribute.InputContext instance and returning a
options tf.distribute.InputOptions used to control options on how this dataset is distributed.

A tf.distribute.DistributedDataset.


View source

Creates tf.distribute.DistributedDataset from

The returned tf.distribute.DistributedDataset can be iterated over similar to regular datasets. NOTE: The user cannot add any more transformations to a tf.distribute.DistributedDataset. You can only create an iterator or examine the tf.TypeSpec of the data generated by it. See API docs of tf.distribute.DistributedDataset to learn more.

The following is an example:

global_batch_size = 2
# Passing the devices is optional.
strategy = tf.distribute.MirroredStrategy(devices=["GPU:0", "GPU:1"])
# Create a dataset
dataset =
# Distribute that dataset
dist_dataset = strategy.experimental_distribute_dataset(dataset)
def replica_fn(input):
  return input*2
result = []
# Iterate over the `tf.distribute.DistributedDataset`
for x in dist_dataset:
  # process dataset elements
  result.append(, args=(x,)))
  0: <tf.Tensor: shape=(1,), dtype=int64, numpy=array([0])>,
  1: <tf.Tensor: shape=(1,), dtype=int64, numpy=array([2])>
}, PerReplica:{
  0: <tf.Tensor: shape=(1,), dtype=int64, numpy=array([4])>,
  1: <tf.Tensor: shape=(1,), dtype=int64, numpy=array([6])>

Three key actions happending under the hood of this method are batching, sharding, and prefetching.

In the code snippet above, dataset is batched by global_batch_size, and calling experimental_distribute_dataset on it rebatches dataset to a new batch size that is equal to the global batch size divided by the number of replicas in sync. We iterate through it using a Pythonic for loop. x is a tf.distribute.DistributedValues containing data for all replicas, and each replica gets data of the new batch size. will take care of feeding the right per-replica data in x to the right replica_fn executed on each replica.

Sharding contains autosharding across multiple workers and within every worker. First, in multi-worker distributed training (i.e. when you use tf.distribute.experimental.MultiWorkerMirroredStrategy or tf.distribute.TPUStrategy), autosharding a dataset over a set of workers means that each worker is assigned a subset of the entire dataset (if the right is set). This is to ensure that at each step, a global batch size of non-overlapping dataset elements will be processed by each worker. Autosharding has a couple of different options that can be specified using Then, sharding within each worker means the method will split the data among all the worker devices (if more than one a present). This will happen regardless of multi-worker autosharding.

By default, this method adds a prefetch transformation at the end of the user provided instance. Th