TPUAccelerator

class lightning.pytorch.accelerators.TPUAccelerator(*args, **kwargs)[source]

Bases: lightning.pytorch.accelerators.accelerator.Accelerator

Accelerator for TPU devices.

Warning

Use of this accelerator beyond import and instantiation is experimental.

static auto_device_count()[source]

Get the devices when set to auto.

Return type

int

get_device_stats(device)[source]

Gets stats for the given TPU device.

Parameters

device (Union[device, str, int]) – TPU device for which to get stats

Return type

Dict[str, Any]

Returns

A dictionary mapping the metrics (free memory and peak memory) to their values.

static get_parallel_devices(devices)[source]

Gets parallel devices for the Accelerator.

Return type

List[int]

static is_available()[source]

Detect if the hardware is available.

Return type

bool

static parse_devices(devices)[source]

Accelerator device parsing logic.

Return type

Union[List[int], int, None]

setup_device(device)[source]

Create and prepare the device for the current process.

Return type

None

teardown()[source]

Clean up any state created by the accelerator.

Return type

None