nntoolbox.optim.lr_scheduler module

class nntoolbox.optim.lr_scheduler.CyclicalTriangularLR(optimizer: torch.optim.optimizer.Optimizer, min_lr: float, max_lr: float, cycle_length: int, inc_fraction: float)

Bases: nntoolbox.optim.lr_scheduler.FunctionalLR

class nntoolbox.optim.lr_scheduler.FunctionalLR(optimizer: torch.optim.optimizer.Optimizer, schedule_fn: Callable[[int], float], last_epoch: int = - 1)

Bases: torch.optim.lr_scheduler.LambdaLR

Calculate learning rate based on a function

get_lr() → List[float]

class nntoolbox.optim.lr_scheduler.TriangularLR(optimizer: torch.optim.optimizer.Optimizer, min_lr: float, max_lr: float, cycle_length: int, inc_fraction: float)

Bases: nntoolbox.optim.lr_scheduler.FunctionalLR