# Copyright (c) Facebook, Inc. and its affiliates.
#
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
from . import LegacyFairseqLRScheduler, register_lr_scheduler
[docs]@register_lr_scheduler("fixed")
class FixedSchedule(LegacyFairseqLRScheduler):
"""Decay the LR on a fixed schedule."""
def __init__(self, args, optimizer):
super().__init__(args, optimizer)
# set defaults
args.warmup_updates = getattr(args, "warmup_updates", 0) or 0
self.lr = args.lr[0]
if args.warmup_updates > 0:
self.warmup_factor = 1.0 / args.warmup_updates
else:
self.warmup_factor = 1
[docs] @staticmethod
def add_args(parser):
"""Add arguments to the parser for this LR scheduler."""
# fmt: off
parser.add_argument('--force-anneal', '--fa', type=int, metavar='N',
help='force annealing at specified epoch (epochs start at 1)')
parser.add_argument('--lr-shrink', default=0.1, type=float, metavar='LS',
help='shrink factor for annealing, lr_new = (lr * lr_shrink)')
parser.add_argument('--warmup-updates', default=0, type=int, metavar='N',
help='warmup the learning rate linearly for the first N updates')
# fmt: on
[docs] def state_dict(self):
return {"lr": self.lr}
[docs] def load_state_dict(self, state_dict):
if "lr" in state_dict:
self.lr = state_dict["lr"]
[docs] def get_next_lr(self, epoch):
lrs = self.args.lr
if self.args.force_anneal is None or epoch < self.args.force_anneal:
# use fixed LR schedule
next_lr = lrs[min(epoch - 1, len(lrs) - 1)]
else:
# annneal based on lr_shrink
next_lr = lrs[-1] * self.args.lr_shrink ** (
epoch + 1 - self.args.force_anneal
)
return next_lr
[docs] def step_begin_epoch(self, epoch):
"""Update the learning rate at the beginning of the given epoch."""
self.lr = self.get_next_lr(epoch)
self.optimizer.set_lr(self.warmup_factor * self.lr)
return self.optimizer.get_lr()
[docs] def step_update(self, num_updates):
"""Update the learning rate after each update."""
if self.args.warmup_updates > 0 and num_updates < self.args.warmup_updates:
self.warmup_factor = (num_updates + 1) / float(self.args.warmup_updates)
self.optimizer.set_lr(self.warmup_factor * self.lr)
else:
self.optimizer.set_lr(self.lr)
return self.optimizer.get_lr()