draugr.torch_utilities.optimisation.scheduling.lr_scheduler.WarmupMultiStepLR¶
- class draugr.torch_utilities.optimisation.scheduling.lr_scheduler.WarmupMultiStepLR(optimiser, milestones, gamma=0.1, warmup_factor=0.3333333333333333, warmup_iters=500, last_epoch=- 1)[source]¶
Bases:
_LRScheduler
description
- __init__(optimiser, milestones, gamma=0.1, warmup_factor=0.3333333333333333, warmup_iters=500, last_epoch=- 1)[source]¶
Methods
__init__
(optimiser, milestones[, gamma, ...])Return last computed learning rate by current scheduler.
get_lr
()- return
load_state_dict
(state_dict)Loads the schedulers state.
print_lr
(is_verbose, group, lr[, epoch])Display the current learning rate.
Returns the state of the scheduler as a
dict
.step
([epoch])- get_last_lr()¶
Return last computed learning rate by current scheduler.
- load_state_dict(state_dict)¶
Loads the schedulers state.
- Parameters
state_dict (dict) – scheduler state. Should be an object returned from a call to
state_dict()
.
- print_lr(is_verbose, group, lr, epoch=None)¶
Display the current learning rate.