xlnstorch.optim.lr_scheduler.LNSMultiplicativeLR

class xlnstorch.optim.lr_scheduler.LNSMultiplicativeLR(optimizer, lr_lambda, last_epoch=-1)

An LNS learning rate scheduler that sets the learning rate of each parameter group to the previous learning rate multipled by a given multiplicative factor.

See also: torch.optim.lr_scheduler.MultiplicativeLR

Parameters:

• optimizer (LNSOptimizer) – Wrapped optimizer.

• lr_lambda (Callable[[int], float | LNSTensor] or List[Callable[[int], float | LNSTensor]]) – A function or a list of functions which computes a multiplicative factor given an integer parameter epoch, which is the index of the current epoch.

• last_epoch (int, optional) – The index of last epoch. Default: -1.

__init__(optimizer, lr_lambda, last_epoch=-1)

Parameters:

• optimizer (LNSOptimizer)

• lr_lambda (Callable[[int], float | LNSTensor] | List[Callable[[int], float | LNSTensor]])

• last_epoch (int)

Methods

__init__(optimizer, lr_lambda[, last_epoch])
get_last_lr()	Return last computed learning rate by current scheduler.
get_lr()	Compute the learning rate of each parameter group.
load_state_dict(state_dict)	Load the scheduler's state.
state_dict()	Return the state of the scheduler as a dict.
step([epoch])	Perform a step.