tf.distribute.experimental.CentralStorageStrategy

A one-machine strategy that puts all variables on a single device.

Inherits From: Strategy

tf.distribute.experimental.CentralStorageStrategy(
    compute_devices=None, parameter_device=None
)

Used in the notebooks

Used in the guide
Distributed training with TensorFlow

Variables are assigned to local CPU or the only GPU. If there is more than one GPU, compute operations (other than variable update operations) will be replicated across all GPUs.

For Example:

strategy = tf.distribute.experimental.CentralStorageStrategy()
# Create a dataset
ds = tf.data.Dataset.range(5).batch(2)
# Distribute that dataset
dist_dataset = strategy.experimental_distribute_dataset(ds)

with strategy.scope():
  @tf.function
  def train_step(val):
    return val + 1

  # Iterate over the distributed dataset
  for x in dist_dataset:
    # process dataset elements
    strategy.run(train_step, args=(x,))

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

Methods

`experimental_assign_to_logical_device`

View source

experimental_assign_to_logical_device(
    tensor, logical_device_id
)

Adds annotation that tensor will be assigned to a logical device.


# Initializing TPU system with 2 logical devices and 4 replicas.
resolver = tf.distribute.cluster_resolver.TPUClusterResolver(tpu='')
tf.config.experimental_connect_to_cluster(resolver)
topology = tf.tpu.experimental.initialize_tpu_system(resolver)
device_assignment = tf.tpu.experimental.DeviceAssignment.build(
    topology,
    computation_shape=[1, 1, 2],
    num_replicas=4)
strategy = tf.distribute.experimental.TPUStrategy(
    resolver, device_assignment=device_assignment)
iterator = iter(inputs)

@tf.function()
def step_fn(inputs):
  output = tf.add(inputs, inputs)

  // Add operation will be executed on logical device 0.
  output = strategy.experimental_assign_to_logical_device(output, 0)
  return output

strategy.run(step_fn, args=(next(iterator),))

Args
`tensor`	Input tensor to annotate.
`logical_device_id`	Id of the logical core to which the tensor will be assigned.

Raises
`ValueError`	The logical device id presented is not consistent with total number of partitions specified by the device assignment.

Returns
Annotated tensor with idential value as `tensor`.

`experimental_distribute_dataset`

View source

experimental_distribute_dataset(
    dataset
)

Distributes a tf.data.Dataset instance provided via dataset.

The returned dataset is a wrapped strategy dataset which creates a multidevice iterator under the hood. It prefetches the input data to the specified devices on the worker. The returned distributed dataset can be iterated over similar to how regular datasets can.

For Example:

strategy = tf.distribute.CentralStorageStrategy()  # with 1 CPU and 1 GPU
dataset = tf.data.Dataset.range(10).batch(2)
dist_dataset = strategy.experimental_distribute_dataset(dataset)
for x in dist_dataset:
  print(x)  # Prints PerReplica values [0, 1], [2, 3],...

Args: dataset: tf.data.Dataset to be prefetched to device.

Returns
A "distributed `Dataset`" that the caller can iterate over.

`experimental_distribute_datasets_from_function`

View source

experimental_distribute_datasets_from_function(
    dataset_fn
)

Distributes tf.data.Dataset instances created by calls to dataset_fn.

dataset_fn will be called once for each worker in the strategy. In this case, we only have one worker so dataset_fn is called once. Each replica on this worker will then dequeue a batch of elements from this local dataset.

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

For Example:

def dataset_fn(input_context):
  batch_size = input_context.get_per_replica_batch_size(global_batch_size)
  d = tf.data.Dataset.from_tensors([[1.]]).repeat().batch(batch_size)
  return d.shard(
      input_context.num_input_pipelines, input_context.input_pipeline_id)

inputs = strategy.experimental_distribute_datasets_from_function(dataset_fn)

for batch in inputs:
  replica_results = strategy.run(replica_fn, args=(batch,))

Args
`dataset_fn`	A function taking a `tf.distribute.InputContext` instance and returning a `tf.data.Dataset`.

Returns
A "distributed `Dataset`", which the caller can iterate over like regular datasets.

`experimental_distribute_values_from_function`

View source

experimental_distribute_values_from_function(
    value_fn
)

Generates tf.distribute.DistributedValues from value_fn.

This function is to generate tf.distribute.DistributedValues to pass into run, reduce, or other methods that take distributed values when not using datasets.

Args
`value_fn`	The function to run to generate values. It is called for each replica with `tf.distribute.ValueContext` as the sole argument. It must return a Tensor or a type that can be converted to a Tensor.

Returns
A `tf.distribute.DistributedValues` containing a value for each replica.

Example usage:

Return constant value per replica:

strategy = tf.distribute.MirroredStrategy()
def value_fn(ctx):
  return tf.constant(1.)
distributed_values = (
     strategy.experimental_distribute_values_from_function(
       value_fn))
local_result = strategy.experimental_local_results(distributed_values)
local_result
(<tf.Tensor: shape=(), dtype=float32, numpy=1.0>,)

Distribute values in array based on replica_id:

strategy = tf.distribute.MirroredStrategy()
array_value = np.array([3., 2., 1.])
def value_fn(ctx):
  return array_value[ctx.replica_id_in_sync_group]
distributed_values = (
     strategy.experimental_distribute_values_from_function(
       value_fn))
local_result = strategy.experimental_local_results(distributed_values)
local_result
(3.0,)

Specify values using num_replicas_in_sync:

strategy = tf.distribute.MirroredStrategy()
def value_fn(ctx):
  return ctx.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,)

Place values on devices and distribute:

strategy = tf.distribute.TPUStrategy()
worker_devices = strategy.extended.worker_devices
multiple_values = []
for i in range(strategy.num_replicas_in_sync):
  with tf.device(worker_devices[i]):
    multiple_values.append(tf.constant(1.0))

def value_fn(ctx):
  return multiple_values[ctx.replica_id]

distributed_values = strategy.
  experimental_distribute_values_from_function(
  value_fn)

`experimental_local_results`

View source

experimental_local_results(
    value
)

Returns the list of all local per-replica values contained in value.

In CentralStorageStrategy there is a single worker so the value returned will be all the values on that worker.

Args
`value`	A value returned by `run()`, `extended.call_for_each_replica()`, or a variable created in `scope`.

Returns
A tuple of values contained in `value`. If `value` represents a single value, this returns `(value,).`

`experimental_make_numpy_dataset`

View source

experimental_make_numpy_dataset(
    numpy_input
)

Makes a tf.data.Dataset for input provided via a numpy array.

This avoids adding numpy_input as a large constant in the graph, and copies the data to the machine or machines that will be processing the input.

Note that you will likely need to use experimental_distribute_dataset with the returned dataset to further distribute it with the strategy.

Example:

numpy_input = np.ones([10], dtype=np.float32)
dataset = strategy.experimental_make_numpy_dataset(numpy_input)
dist_dataset = strategy.experimental_distribute_dataset(dataset)

Args
`numpy_input`	A nest of NumPy input arrays that will be converted into a dataset. Note that lists of Numpy arrays are stacked, as that is normal `tf.data.Dataset` behavior.

Returns
A `tf.data.Dataset` representing `numpy_input`.

`experimental_replicate_to_logical_devices`

View source

experimental_replicate_to_logical_devices(
    tensor
)

Adds annotation that tensor will be replicated to all logical devices.

# Initializing TPU system with 2 logical devices and 4 replicas.
resolver = tf.distribute.cluster_resolver.TPUClusterResolver(tpu='')
tf.config.experimental_connect_to_cluster(resolver)
topology = tf.tpu.experimental.initialize_tpu_system(resolver)
device_assignment = tf.tpu.experimental.DeviceAssignment.build(
    topology,
    computation_shape=[1, 1, 2],
    num_replicas=4)
strategy = tf.distribute.experimental.TPUStrategy(
    resolver, device_assignment=device_assignment)

iterator = iter(inputs)

@tf.function()
def step_fn(inputs):
  images, labels = inputs
  images = strategy.experimental_split_to_logical_devices(
    inputs, [1, 2, 4, 1])

  // model() function will be executed on 8 logical devices with `inputs`
  // split 2 * 4  ways.
  output = model(inputs)

  // For loss calculation, all logical devices share the same logits
  // and labels.
  labels = strategy.experimental_replicate_to_logical_devices(labels)
  output = strategy.experimental_replicate_to_logical_devices(output)
  loss = loss_fn(labels, output)

  return loss

strategy.run(step_fn, args=(next(iterator),))

Args: tensor: Input tensor to annotate.

Returns
Annotated tensor with idential value as `tensor`.

`experimental_split_to_logical_devices`

View source

experimental_split_to_logical_devices(
    tensor, partition_dimensions
)

Adds annotation that tensor will be split across logical devices.

For example, for system with 8 logical devices, if tensor is an image tensor with shape (batch_size, width, height, channel) and partition_dimensions is [1, 2, 4, 1], then tensor will be split 2 in width dimension and 4 way in height dimension and the split tensor values will be fed into 8 logical devices.

# Initializing TPU system with 8 logical devices and 1 replica.
resolver = tf.distribute.cluster_resolver.TPUClusterResolver(tpu='')
tf.config.experimental_connect_to_cluster(resolver)
topology = tf.tpu.experimental.initialize_tpu_system(resolver)
device_assignment = tf.tpu.experimental.DeviceAssignment.build(
    topology,
    computation_shape=[2, 2, 2],
    num_replicas=1)
strategy = tf.distribute.experimental.TPUStrategy(
    resolver, device_assignment=device_assignment)

iterator = iter(inputs)

@tf.function()
def step_fn(inputs):
  inputs = strategy.experimental_split_to_logical_devices(
    inputs, [1, 2, 4, 1])

  // model() function will be executed on 8 logical devices with `inputs`
  // split 2 * 4  ways.
  output = model(inputs)
  return output

strategy.run(step_fn, args=(next(iterator),))

Args: tensor: Input tensor to annotate. partition_dimensions: An unnested list of integers with the size equal to rank of tensor specifying how tensor will be partitioned. The product of all elements in partition_dimensions must be equal to the total number of logical devices per replica.

Raises

Raises
`ValueError`	1) If the size of partition_dimensions does not equal to rank of `tensor` or 2) if product of elements of `partition_dimensions` does not match the number of logical devices per replica defined by the implementing DistributionStrategy's device specification or 3) if a known size of `tensor` is not divisible by corresponding value in `partition_dimensions`.

ValueError

1) If the size of partition_dimensions does not equal to rank of tensor or 2) if product of elements of partition_dimensions does not match the number of logical devices per replica defined by the implementing DistributionStrategy's device specification or 3) if a known size of tensor is not divisible by corresponding value in partition_dimensions.

Returns
Annotated tensor with idential value as `tensor`.

`reduce`

View source

reduce(
    reduce_op, value, axis
)

Reduce value across replicas.

Given a per-replica value returned by run, say a per-example loss, the batch will be divided across all the replicas. This function allows you to aggregate across replicas and optionally also across batch elements. For example, if you have a global batch size of 8 and 2 replicas, values for examples [0, 1, 2, 3] will be on replica 0 and [4, 5, 6, 7] will be on replica 1. By default, reduce will just aggregate across replicas, returning [0+4, 1+5, 2+6, 3+7]. This is useful when each replica is computing a scalar or some other value that doesn't have a "batch" dimension (like a gradient). More often you will want to aggregate across the global batch, which you can get by specifying the batch dimension as the axis, typically axis=0. In this case it would return a scalar 0+1+2+3+4+5+6+7.

If there is a last partial batch, you will need to specify an axis so that the resulting shape is consistent across replicas. So if the last batch has size 6 and it is divided into [0, 1, 2, 3] and [4, 5], you would get a shape mismatch unless you specify axis=0. If you specify tf.distribute.ReduceOp.MEAN, using axis=0 will use the correct denominator of 6. Contrast this with computing reduce_mean to get a scalar value on each replica and this function to average those means, which will weigh some values 1/8 and others 1/4.

For Example:

strategy = tf.distribute.experimental.CentralStorageStrategy(
    compute_devices=['CPU:0', 'GPU:0'], parameter_device='CPU:0')
ds = tf.data.Dataset.range(10)
# Distribute that dataset
dist_dataset = strategy.experimental_distribute_dataset(ds)

with strategy.scope():
  @tf.function
  def train_step(val):
    # pass through
    return val

  # Iterate over the distributed dataset
  for x in dist_dataset:
    result = strategy.run(train_step, args=(x,))

result = strategy.reduce(tf.distribute.ReduceOp.SUM, result,
                         axis=None).numpy()
# result: array([ 4,  6,  8, 10])

result = strategy.reduce(tf.distribute.ReduceOp.SUM, result, axis=0).numpy()
# result: 28

Args
`reduce_op`	A `tf.distribute.ReduceOp` value specifying how values should be combined.
`value`	A "per replica" value, e.g. returned by `run` to be combined into a single tensor.
`axis`	Specifies the dimension to reduce along within each replica's tensor. Should typically be set to the batch dimension, or `None` to only reduce across replicas (e.g. if the tensor has no batch dimension).

Returns
A `Tensor`.

`run`

View source

run(
    fn, args=(), kwargs=None, options=None
)

Run fn on each replica, with the given arguments.

In CentralStorageStrategy, fn is called on each of the compute replicas, with the provided "per replica" arguments specific to that device.

Args
`fn`	The function to run. The output must be a `tf.nest` of `Tensor`s.
`args`	(Optional) Positional arguments to `fn`.
`kwargs`	(Optional) Keyword arguments to `fn`.
`options`	(Optional) An instance of `tf.distribute.RunOptions` specifying the options to run `fn`.

Returns
Return value from running `fn`.

`scope`

View source

scope()

Returns a context manager selecting this Strategy as current.

Inside a with strategy.scope(): code block, this thread will use a variable creator set by strategy, and will enter its "cross-replica context".

Returns
A context manager.

tf.distribute.experimental.CentralStorageStrategy

Used in the notebooks

For Example:

Attributes

Methods

experimental_assign_to_logical_device

experimental_distribute_dataset

For Example:

experimental_distribute_datasets_from_function

For Example:

experimental_distribute_values_from_function

Example usage:

experimental_local_results

experimental_make_numpy_dataset

Example:

experimental_replicate_to_logical_devices

experimental_split_to_logical_devices

reduce

For Example:

run

scope

`experimental_assign_to_logical_device`

`experimental_distribute_dataset`

`experimental_distribute_datasets_from_function`

`experimental_distribute_values_from_function`

`experimental_local_results`

`experimental_make_numpy_dataset`

`experimental_replicate_to_logical_devices`

`experimental_split_to_logical_devices`

`reduce`

`run`

`scope`