SingleDevicePlugin¶

class pytorch_lightning.plugins.training_type.SingleDevicePlugin(device)[source]¶

Bases: pytorch_lightning.plugins.training_type.training_type_plugin.TrainingTypePlugin

Plugin that handles communication on a single device.

all_gather(tensor, group=None, sync_grads=False)[source]¶

Perform a all_gather on all processes

Return type: Tensor

barrier(*args, **kwargs)[source]¶

Forces all possibly joined processes to wait for each other

Return type: None

broadcast(obj, src=0)[source]¶

Broadcasts an object to all processes

Return type: object

model_to_device()[source]¶

Moves the model to the correct device

Return type: None

reduce(tensor, *args, **kwargs)[source]¶

Reduces a tensor from several distributed processes to one aggregated tensor. As this plugin only operates with a single device, the reduction is simply the identity.

Parameters

tensor¶ (Union[Any, Tensor]) – the tensor to sync and reduce
*args¶ – ignored
**kwargs¶ – ignored

Return type

Union[Any, Tensor]

Returns

the unmodified input as reduction is not needed for single process operation

setup(model)[source]¶

Called by the accelerator to finish setup.

Return type: Module

property is_global_zero¶

Whether the current process is the rank zero process not only on the local node, but for all nodes.

Return type: bool

property on_gpu¶

Returns whether the current process is done on GPU

Return type: bool

property root_device¶

Returns the root device

Return type: device