"""Warm up learning rate scheduler module."""
from typing import Union
import torch
from torch.optim.lr_scheduler import _LRScheduler
from typeguard import check_argument_types
from espnet2.schedulers.abs_scheduler import AbsBatchStepScheduler
[docs]class WarmupLR(_LRScheduler, AbsBatchStepScheduler):
"""The WarmupLR scheduler
This scheduler is almost same as NoamLR Scheduler except for following difference:
NoamLR:
lr = optimizer.lr * model_size ** -0.5
* min(step ** -0.5, step * warmup_step ** -1.5)
WarmupLR:
lr = optimizer.lr * warmup_step ** 0.5
* min(step ** -0.5, step * warmup_step ** -1.5)
Note that the maximum lr equals to optimizer.lr in this scheduler.
"""
def __init__(
self,
optimizer: torch.optim.Optimizer,
warmup_steps: Union[int, float] = 25000,
last_epoch: int = -1,
):
assert check_argument_types()
self.warmup_steps = warmup_steps
# __init__() must be invoked before setting field
# because step() is also invoked in __init__()
super().__init__(optimizer, last_epoch)
def __repr__(self):
return f"{self.__class__.__name__}(warmup_steps={self.warmup_steps})"
[docs] def get_lr(self):
step_num = self.last_epoch + 1
return [
lr
* self.warmup_steps**0.5
* min(step_num**-0.5, step_num * self.warmup_steps**-1.5)
for lr in self.base_lrs
]