TorchScheduler#

class TorchScheduler(scheduler_class, **kwargs)[source]#

Bases: Scheduler

A wrapper class for using PyTorch schedulers.

Initialization of the TorchScheduler class.

Parameters:

  • scheduler_class (torch.optim.LRScheduler) – A torch.optim.LRScheduler class.

  • kwargs (dict) – Additional parameters passed to scheduler_class, see more here <https://pytorch.org/docs/stable/optim.html#algorithms>_.

hook(optimizer)[source]#

Initialize the scheduler instance with the given parameters.

Parameters:

parameters (dict) – The parameters of the optimizer.

property instance#

Get the scheduler instance.

Returns:

The scheduelr instance.

Return type:

torch.optim.LRScheduler