DataParallelPlugin

class pytorch_lightning.plugins.training_type.DataParallelPlugin(parallel_devices=None, checkpoint_io=None)

Bases: pytorch_lightning.plugins.training_type.parallel.ParallelPlugin

Implements data-parallel training in a single process, i.e., the model gets replicated to each device and each gets a split of the data.

barrier(*args, **kwargs)

Synchronizes all processes which blocks processes until the whole group enters this function.

Parameters

name – an optional name to pass into barrier.

broadcast(obj, src=0)

Broadcasts an object to all processes.

Parameters

• obj (object) – the object to broadcast

• src (int) – source rank

Return type

object

model_to_device()

Moves the model to the correct device.

Return type

None

reduce(collection, *args, **kwargs)

Reduces a collection of tensors from all processes. It can be applied to just a single tensor.

Parameters

• collection (Union[Metric, Tensor, int, float, Mapping[str, Union[Metric, Tensor, int, float]]]) – The collection of tensors to sync and reduce.

• *args – ignored for DP

• **kwargs – ignored for DP

Return type

Union[Metric, Tensor, int, float, Mapping[str, Union[Metric, Tensor, int, float]]]

Returns

Reduced tensor values or the same value if it was not or did not contain a tensor.

reduce_boolean_decision(decision)

Reduce the early stopping decision across all processes.

Return type

bool

setup()

Called by the accelerator to finish setup.

Return type

None

teardown()

This method is called to teardown the training process.

It is the right place to release memory and free other resources.

Return type

None

property root_device

Return the root device.