SingleHPUStrategy¶

class pytorch_lightning.strategies.SingleHPUStrategy(device='hpu', accelerator=None, checkpoint_io=None, precision_plugin=None)[source]¶

Bases: pytorch_lightning.strategies.single_device.SingleDeviceStrategy

Strategy for training on single HPU device.

model_to_device()[source]¶

Moves the model to the correct device.

Return type:: None

optimizer_step(optimizer, opt_idx, closure, model=None, **kwargs)[source]¶

Performs the actual optimizer step.

Parameters:

optimizer¶ (Optimizer) – the optimizer performing the step
opt_idx¶ (int) – index of the current optimizer
closure¶ (Callable[[], Any]) – closure calculating the loss value
model¶ (Union[LightningModule, Module, None]) – reference to the model, optionally defining optimizer step related hooks
**kwargs¶ (Any) – Keyword arguments to to optimizer.step

Return type:

Any

setup(trainer)[source]¶

Setup plugins for the trainer fit and creates optimizers.

Parameters:: trainer¶ (Trainer) – the trainer instance
Return type:: None

setup_optimizers(trainer)[source]¶

Creates optimizers and schedulers.

Parameters:: trainer¶ (Trainer) – the Trainer, these optimizers should be connected to
Return type:: None