xlnstorch.optim.lr_scheduler.LNSLinearLR

class xlnstorch.optim.lr_scheduler.LNSLinearLR(optimizer, start_factor=0.3333333333333333, end_factor=1.0, total_iters=0, last_epoch=-1)

An LNS learning rate scheduler that sets the learning rate of each parameter group to a linearly decaying value.

See also: torch.optim.lr_scheduler.LinearLR

Parameters:

• optimizer (LNSOptimizer) – Wrapped optimizer.

• start_factor (float | LNSTensor) – The initial factor for the learning rate.

• end_factor (float | LNSTensor) – The final factor for the learning rate.

• total_iters (int, optional) – The number of iterations over which the learning rate will decay. Default: 0.

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

__init__(optimizer, start_factor=0.3333333333333333, end_factor=1.0, total_iters=0, last_epoch=-1)

Parameters:

• optimizer (LNSOptimizer)

• start_factor (float | LNSTensor)

• end_factor (float | LNSTensor)

• total_iters (int)

• last_epoch (int)

Methods

__init__(optimizer[, start_factor, ...])
get_last_lr()	Return last computed learning rate by current scheduler.
get_lr()	Compute the learning rate.
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.