# Module: tf.distribute.experimental

<br />

Experimental Distribution Strategy library.

## Modules

[`coordinator`](https://www.tensorflow.org/versions/r2.14/api_docs/python/tf/distribute/experimental/coordinator) module: Public API for tf.distribute.experimental.coordinator namespace.

[`partitioners`](https://www.tensorflow.org/versions/r2.14/api_docs/python/tf/distribute/experimental/partitioners) module: Public API for tf.distribute.experimental.partitioners namespace.

[`rpc`](https://www.tensorflow.org/versions/r2.14/api_docs/python/tf/distribute/experimental/rpc) module: Public API for tf.distribute.experimental.rpc namespace.

## Classes

[`class CentralStorageStrategy`](https://www.tensorflow.org/versions/r2.14/api_docs/python/tf/distribute/experimental/CentralStorageStrategy): A one-machine strategy that puts all variables on a single device.

[`class CollectiveCommunication`](https://www.tensorflow.org/versions/r2.14/api_docs/python/tf/distribute/experimental/CommunicationImplementation): Cross device communication implementation.

[`class CollectiveHints`](https://www.tensorflow.org/versions/r2.14/api_docs/python/tf/distribute/experimental/CollectiveHints): Hints for collective operations like AllReduce.

[`class CommunicationImplementation`](https://www.tensorflow.org/versions/r2.14/api_docs/python/tf/distribute/experimental/CommunicationImplementation): Cross device communication implementation.

[`class CommunicationOptions`](https://www.tensorflow.org/versions/r2.14/api_docs/python/tf/distribute/experimental/CommunicationOptions): Options for cross device communications like All-reduce.

[`class MultiWorkerMirroredStrategy`](https://www.tensorflow.org/versions/r2.14/api_docs/python/tf/distribute/experimental/MultiWorkerMirroredStrategy): A distribution strategy for synchronous training on multiple workers.

[`class ParameterServerStrategy`](https://www.tensorflow.org/versions/r2.14/api_docs/python/tf/distribute/experimental/ParameterServerStrategy): An multi-worker tf.distribute strategy with parameter servers.

[`class PreemptionCheckpointHandler`](https://www.tensorflow.org/versions/r2.14/api_docs/python/tf/distribute/experimental/PreemptionCheckpointHandler): Preemption and error handler for synchronous training.

[`class PreemptionWatcher`](https://www.tensorflow.org/versions/r2.14/api_docs/python/tf/distribute/experimental/PreemptionWatcher): Watch preemption signal and store it.

[`class TPUStrategy`](https://www.tensorflow.org/versions/r2.14/api_docs/python/tf/distribute/experimental/TPUStrategy): Synchronous training on TPUs and TPU Pods.

[`class TerminationConfig`](https://www.tensorflow.org/versions/r2.14/api_docs/python/tf/distribute/experimental/TerminationConfig): Customization of `PreemptionCheckpointHandler` for various platforms.

[`class ValueContext`](https://www.tensorflow.org/versions/r2.14/api_docs/python/tf/distribute/experimental/ValueContext): A class wrapping information needed by a distribute function.