paddlespeech.vector.training.scheduler module
- class paddlespeech.vector.training.scheduler.CyclicLRScheduler(base_lr: float = 1e-08, max_lr: float = 0.001, step_size: int = 10000)[source]
Bases:
LRScheduler
Methods
__call__
()Return lastest computed learning rate on current epoch.
get_lr
()For those subclass who overload
LRScheduler
(Base Class), User should have a custom implementation ofget_lr()
.set_dict
(state_dict)Loads the schedulers state.
set_state_dict
(state_dict)Loads the schedulers state.
state_dict
()Returns the state of the scheduler as a
dict
.state_keys
()For those subclass who overload
LRScheduler
(Base Class).step
()step
should be called afteroptimizer.step
.- get_lr()[source]
For those subclass who overload
LRScheduler
(Base Class), User should have a custom implementation ofget_lr()
.Otherwise, an
NotImplementedError
exception will be thrown.
- step()[source]
step
should be called afteroptimizer.step
. It will update the learning rate in optimizer according to currentepoch
. The new learning rate will take effect on nextoptimizer.step
.- Args:
epoch (int, None): specify current epoch. Default: None. Auto-increment from last_epoch=-1.
- Returns:
None