WeHub Mirror
88 lines
3.0 KiB
Python
88 lines
3.0 KiB
Python
# Copyright (c) 2026 PaddlePaddle Authors. 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 __future__ import annotations
|
|
|
|
import warnings
|
|
from typing import TYPE_CHECKING
|
|
|
|
from paddle.optimizer import Adagrad as PaddleAdagrad
|
|
|
|
if TYPE_CHECKING:
|
|
from collections.abc import Callable, Sequence
|
|
|
|
from paddle import Tensor
|
|
from paddle.optimizer.adagrad import _AdagradParameterConfig
|
|
|
|
|
|
class Adagrad(PaddleAdagrad):
|
|
def __init__(
|
|
self,
|
|
params: Sequence[Tensor] | Sequence[_AdagradParameterConfig] | None,
|
|
lr: float | Tensor = 1e-2,
|
|
lr_decay: float = 0,
|
|
weight_decay: float = 0,
|
|
initial_accumulator_value: float = 0,
|
|
eps: float = 1e-10,
|
|
foreach: bool | None = None,
|
|
*,
|
|
maximize: bool = False,
|
|
differentiable: bool = False,
|
|
fused: bool | None = None,
|
|
) -> None:
|
|
if foreach is not None or differentiable is True or fused is not None:
|
|
warnings.warn(
|
|
"foreach, differentiable, fused are currently not supported in Adagrad and will be ignored. "
|
|
"The parameters are reserved for future implementation."
|
|
)
|
|
self._lr_decay = None
|
|
if lr_decay != 0.0:
|
|
self._lr_decay = lr_decay
|
|
self._step = -1
|
|
super().__init__(
|
|
learning_rate=lr,
|
|
epsilon=eps,
|
|
parameters=params,
|
|
weight_decay=weight_decay,
|
|
initial_accumulator_value=initial_accumulator_value,
|
|
maximize=maximize,
|
|
)
|
|
|
|
def state_dict(self) -> dict[str, Tensor]:
|
|
state_dict = super().state_dict()
|
|
if self._lr_decay is not None:
|
|
state_dict['step'] = self._step
|
|
return state_dict
|
|
|
|
def set_state_dict(self, state_dict: dict[str, Tensor]) -> None:
|
|
state_dict = state_dict.copy()
|
|
if "step" in state_dict:
|
|
if self._lr_decay is not None:
|
|
self._step = state_dict["step"]
|
|
state_dict.pop("step")
|
|
return super().set_state_dict(state_dict)
|
|
|
|
def _create_param_lr(self, param_and_grad):
|
|
param_lr = super()._create_param_lr(param_and_grad)
|
|
if self._lr_decay is not None:
|
|
param_lr = param_lr / (1.0 + self._step * self._lr_decay)
|
|
return param_lr
|
|
|
|
def step(
|
|
self, closure: Callable[[], Tensor] | None = None
|
|
) -> Tensor | None:
|
|
if self._lr_decay is not None:
|
|
self._step += 1
|
|
return super().step(closure)
|