# Copyright 2021 cstsunfu. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from typing import Dict
import math
from dlk.utils.config import BaseConfig
from . import scheduler_register, scheduler_config_register, BaseScheduler
from torch.optim.lr_scheduler import LambdaLR
import torch.optim as optim
[docs]@scheduler_config_register("cosine_warmup")
class CosineWarmupScheduleConfig(BaseConfig):
"""Config for CosineWarmupSchedule
Config Example:
>>> {
>>> "config": {
>>> "last_epoch": -1,
>>> "num_warmup_steps": 0,
>>> "num_training_steps": -1,
>>> "num_cycles": 0.5,
>>> },
>>> "_name": "cosine_warmup",
>>> }
"""
def __init__(self, config: Dict):
super(CosineWarmupScheduleConfig, self).__init__(config)
config = config['config']
self.last_epoch = config["last_epoch"]
self.num_warmup_steps = config["num_warmup_steps"]
self.num_training_steps = config["num_training_steps"]
self.num_cycles = config['num_cycles']
self.post_check(config, used=[
"last_epoch",
"num_warmup_steps",
"num_training_steps",
"num_cycles",
])
[docs]@scheduler_register("cosine_warmup")
class CosineWarmupSchedule(BaseScheduler):
"""CosineWarmupSchedule"""
def __init__(self, optimizer: optim.Optimizer, config: CosineWarmupScheduleConfig):
super(CosineWarmupSchedule, self).__init__()
self.config = config
self.optimizer = optimizer
[docs] def get_scheduler(self)->LambdaLR:
"""return the initialized linear wramup then cos decay scheduler
Returns:
Schedule
"""
num_training_steps = self.config.num_training_steps
num_warmup_steps = self.config.num_warmup_steps
if num_warmup_steps >0 and num_warmup_steps < 1:
num_warmup_steps = int(num_warmup_steps * num_training_steps)
last_epoch = self.config.last_epoch
num_cycles = self.config.num_cycles
def lr_lambda(current_step):
if current_step < num_warmup_steps:
return float(current_step) / float(max(1, num_warmup_steps))
progress = float(current_step - num_warmup_steps) / float(max(1, num_training_steps - num_warmup_steps))
return max(0.0, 0.5 * (1.0 + math.cos(math.pi * float(num_cycles) * 2.0 * progress)))
return LambdaLR(self.optimizer, lr_lambda, last_epoch)