# Copyright 2023 Huawei Technologies Co., Ltd
#
# 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.
# ============================================================================
"""rmsprop"""
from __future__ import absolute_import
from mindspore.ops import functional as F, composite as C, operations as P
import mindspore.common.dtype as mstype
from mindspore.experimental.optim.optimizer import Optimizer, check_not_less_than, check_not_less_than_without_equal
from mindspore import ops
from mindspore import jit
_rmsprop_opt = C.MultitypeFuncGraph("rmsprop_opt")
op_mul = P.Mul()
op_sqrt = P.Sqrt()
@_rmsprop_opt.register("Bool", "Number", "Number", "Number", "Tensor", "Tensor", "Tensor", "Tensor", "Tensor", "Tensor")
def _run_rmsprop_opt(centered, alpha, eps, momentum, lr, weight, mean_square, mean_grad, mom, grad):
"""Apply rmsprop optimizer to the weight parameter using dynamic learning rate."""
F.assign(mean_square, ops.addcmul(op_mul(mean_square, alpha), grad, grad, 1 - alpha))
if centered:
F.assign(mean_grad, op_mul(mean_grad, alpha) + op_mul(grad, 1 - alpha))
avg = op_sqrt(ops.addcmul(mean_square, mean_grad, mean_grad, -1.)) + eps
else:
avg = op_sqrt(mean_square) + eps
if momentum > 0:
F.assign(mom, op_mul(mom, momentum) + grad / avg)
F.assign(weight, weight - mom * lr)
else:
F.assign(weight, weight - lr * grad / avg)
return True
[docs]class RMSprop(Optimizer):
r"""
Implements RMSprop algorithm.
.. warning::
This is an experimental optimizer API that is subject to change.
This module must be used with lr scheduler module in `LRScheduler Class
<https://www.mindspore.cn/docs/en/r2.3.0rc2/api_python/mindspore.experimental.html#lrscheduler-class>`_ .
Args:
params (Union[list(Parameter), list(dict)]): list of parameters to optimize or dicts defining
parameter groups.
lr (Union[int, float, Tensor], optional): learning rate. Default: ``1e-2``.
alpha (float, optional): smoothing constant. Default: ``0.99``.
eps (float, optional): term added to the denominator to improve numerical stability. Default: ``1e-8``.
weight_decay (float, optional): weight decay (L2 penalty). Default: ``0.``.
momentum (float, optional): momentum factor. Default: ``0.``.
centered (bool, optional) : if ``True``, compute the centered RMSProp,
the gradient is normalized by an estimation of its variance. Default: ``False``.
maximize (bool, optional): maximize the params based on the objective, instead of minimizing.
Default: ``False``.
Inputs:
- **gradients** (tuple[Tensor]) - The gradients of `params`.
Raises:
ValueError: If the learning rate is not int, float or Tensor.
ValueError: If the learning rate is less than 0.
ValueError: If the `momentum` is less than 0.0.
ValueError: If the `alpha` is less than 0.0.
ValueError: If the `eps` is less than 0.0.
ValueError: If the `weight_decay` is less than 0.
Supported Platforms:
``Ascend`` ``GPU`` ``CPU``
Examples:
>>> import mindspore
>>> from mindspore import nn
>>> from mindspore.experimental import optim
>>> # Define the network structure of LeNet5. Refer to
>>> # https://gitee.com/mindspore/docs/blob/r2.3.0rc2/docs/mindspore/code/lenet.py
>>> net = LeNet5()
>>> loss_fn = nn.SoftmaxCrossEntropyWithLogits(sparse=True)
>>> optimizer = optim.RMSprop(net.trainable_params(), lr=0.1)
>>> def forward_fn(data, label):
... logits = net(data)
... loss = loss_fn(logits, label)
... return loss, logits
>>> grad_fn = mindspore.value_and_grad(forward_fn, None, optimizer.parameters, has_aux=True)
>>> def train_step(data, label):
... (loss, _), grads = grad_fn(data, label)
... optimizer(grads)
... return loss
"""
def __init__(self, params, lr=1e-2, alpha=0.99, eps=1e-8, weight_decay=0.0, momentum=0.0,
centered=False, maximize=False):
check_not_less_than_without_equal(lr, "lr", self.cls_name)
check_not_less_than(alpha, "alpha", self.cls_name)
check_not_less_than_without_equal(eps, "eps", self.cls_name)
check_not_less_than(momentum, "momentum", self.cls_name)
check_not_less_than(weight_decay, "weight_decay", self.cls_name)
defaults = dict(
lr=lr,
momentum=momentum,
alpha=alpha,
eps=eps,
centered=centered,
weight_decay=weight_decay,
maximize=maximize,
)
super(RMSprop, self).__init__(params, defaults)
self.mean_grad = self.parameters.clone(prefix="mean_grad", init='zeros')
self.mean_square = self.parameters.clone(prefix="mean_square", init='zeros')
self.moment = self.parameters.clone(prefix="moment", init='zeros')
self.op_cast = P.Cast()
@jit
def implementation(self, group_id, lr, gradients, maximize, weight_decay, centered, alpha, eps, momentum):
"""Extract the common computing part for acceleration"""
start_id = self.group_start_id[group_id]
end_id = self.group_start_id[group_id + 1]
params = self.parameters[start_id: end_id]
grads = tuple([grad if not maximize else F.neg(grad) for grad in gradients[start_id: end_id]])
grads = self._decay_weight(weight_decay, params, grads)
mean_grad = self.mean_grad[start_id: end_id]
mean_square = self.mean_square[start_id: end_id]
moment = self.moment[start_id: end_id]
self.hyper_map(F.partial(_rmsprop_opt, centered, alpha, eps, momentum, lr),
params, mean_square, mean_grad, moment, grads)
return True
def construct(self, gradients):
for group_id, group in enumerate(self.param_groups):
lr = self.lrs[group_id]
if isinstance(group.get("lr"), float):
lr = self.op_cast(group.get("lr"), mstype.float32)
maximize = group.get("maximize")
self.implementation(group_id, lr, gradients, maximize, group["weight_decay"], group["centered"],
group["alpha"], group["eps"],
group["momentum"])
return True