mindspore.mint.optim.adam 源代码

# Copyright 2024 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.
# ============================================================================
"""Adam"""
from __future__ import absolute_import

from mindspore.ops import functional as F, composite as C, operations as P
from mindspore.common.parameter import Parameter
from mindspore.common.tensor import Tensor
from mindspore.common import dtype as mstype
from mindspore.experimental.optim.optimizer import Optimizer
from mindspore import _checkparam as validator
from mindspore import mint

_optim_adamw_opt = C.MultitypeFuncGraph("optim_adamw_opt")
hyper_map = C.HyperMap()
assign_add = P.AssignAdd()

@_optim_adamw_opt.register("Float", "Float", "Float", "Tensor", "Tensor", "Tensor", "Tensor",
                           "Tensor", "Tensor", "Tensor")
def _run_optim_adamw_amsgrad_opt(beta1, beta2, eps, neg_step_size, sqrt_bias_correction2, parameters, grads, exp_avg,
                                 exp_avg_sq, max_exp_avg_sq):
    """Apply adam optimizer to the weight parameter when amsgrad is True."""
    success = True
    #lerp(grads, exp_avg, beta1)
    exp_avg_tmp = mint.add(mint.mul(exp_avg, beta1), grads, alpha=1 - beta1)
    exp_avg_sq_tmp = mint.mul(exp_avg_sq, beta2) + mint.mul(mint.mul(grads, grads), 1 - beta2)

    max_exp_avg_sq = mint.maximum(max_exp_avg_sq, exp_avg_sq_tmp)
    denom = F.cast(mint.div(mint.sqrt(max_exp_avg_sq), sqrt_bias_correction2), max_exp_avg_sq.dtype)
    denom = mint.add(denom, eps)

    delta_param = mint.mul(F.cast(neg_step_size, max_exp_avg_sq.dtype), mint.div(exp_avg_tmp, denom))
    F.assign(exp_avg, exp_avg_tmp)
    F.assign(exp_avg_sq, exp_avg_sq_tmp)
    assign_add(parameters, delta_param)
    return success

@_optim_adamw_opt.register("Float", "Float", "Float", "Tensor", "Tensor", "Tensor", "Tensor", "Tensor", "Tensor")
def _run_optim_adamw_opt(beta1, beta2, eps, neg_step_size, sqrt_bias_correction2, parameters, grads, exp_avg,
                         exp_avg_sq):
    """Apply adam optimizer to the weight parameter when amsgrad is False."""
    success = True
    #lerp(grads, exp_avg, beta1)
    exp_avg_tmp = mint.add(mint.mul(exp_avg, beta1), grads, alpha=1 - beta1)
    exp_avg_sq_tmp = mint.mul(exp_avg_sq, beta2) + mint.mul(mint.mul(grads, grads), 1 - beta2)

    denom = F.cast(mint.div(mint.sqrt(exp_avg_sq_tmp), sqrt_bias_correction2), exp_avg_sq_tmp.dtype)
    denom = mint.add(denom, eps)

    delta_param = mint.mul(F.cast(neg_step_size, exp_avg_sq_tmp.dtype), mint.div(exp_avg_tmp, denom))
    F.assign(exp_avg, exp_avg_tmp)
    F.assign(exp_avg_sq, exp_avg_sq_tmp)
    assign_add(parameters, delta_param)
    return success


def _check_param_value(betas, eps, weight_decay, lr, amsgrad, maximize, prim_name):
    """Check the type of inputs."""
    validator.check_value_type('betas', betas, [tuple], prim_name)
    validator.check("betas size", len(betas), "", [2], validator.IN, prim_name)
    validator.check_value_type("betas[0]", betas[0], [float], prim_name)
    validator.check_value_type("betas[1]", betas[1], [float], prim_name)
    validator.check_value_type("eps", eps, [float], prim_name)
    validator.check_value_type("weight_decay", weight_decay, [float], prim_name)
    validator.check_value_type("lr", lr, [float], prim_name)
    validator.check_value_type("amsgrad", amsgrad, [bool], prim_name)
    validator.check_value_type("maximize", maximize, [bool], prim_name)


[文档]class Adam(Optimizer): r""" Implements Adaptive Moment Estimation (Adam) algorithm. The updating formulas are as follows: .. math:: \begin{aligned} &\textbf{input} : \gamma \text{ (lr)}, \beta_1, \beta_2 \text{ (betas)},\theta_0 \text{ (params)},f(\theta) \text{ (objective)} \\ &\hspace{13mm} \lambda \text{ (weight decay)}, \: \textit{amsgrad}, \:\textit{maximize} \\ &\textbf{initialize} : m_0 \leftarrow 0 \text{ ( first moment)}, v_0\leftarrow 0 \text{ (second moment)},\: \widehat{v_0}^{max}\leftarrow 0\\[-1.ex] &\textbf{for} \: t=1 \: \textbf{to} \: \ldots \: \textbf{do} \\ &\hspace{5mm}\textbf{if} \: \textit{maximize}: \\ &\hspace{10mm}g_t \leftarrow -\nabla_{\theta} f_t (\theta_{t-1}) \\ &\hspace{5mm}\textbf{else} \\ &\hspace{10mm}g_t \leftarrow \nabla_{\theta} f_t (\theta_{t-1}) \\ &\hspace{5mm}\textbf{if} \: \lambda \neq 0 \\ &\hspace{10mm} g_t \leftarrow g_t + \lambda \theta_{t-1} \\ &\hspace{5mm}m_t \leftarrow \beta_1 m_{t-1} + (1 - \beta_1) g_t \\ &\hspace{5mm}v_t \leftarrow \beta_2 v_{t-1} + (1-\beta_2) g^2_t \\ &\hspace{5mm}\widehat{m_t} \leftarrow m_t/\big(1-\beta_1^t \big) \\ &\hspace{5mm}\widehat{v_t} \leftarrow v_t/\big(1-\beta_2^t \big) \\ &\hspace{5mm}\textbf{if} \: amsgrad \\ &\hspace{10mm}\widehat{v_t}^{max} \leftarrow \mathrm{max}(\widehat{v_t}^{max}, \widehat{v_t}) \\ &\hspace{10mm}\theta_t \leftarrow \theta_{t-1} - \gamma \widehat{m_t}/ \big(\sqrt{\widehat{v_t}^{max}} + \epsilon \big) \\ &\hspace{5mm}\textbf{else} \\ &\hspace{10mm}\theta_t \leftarrow \theta_{t-1} - \gamma \widehat{m_t}/ \big(\sqrt{\widehat{v_t}} + \epsilon \big) \\ &\bf{return} \: \theta_t \\[-1.ex] \end{aligned} .. warning:: This is an experimental API that is subject to change or deletion. 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-3``. betas (Tuple[float, float], optional): The exponential decay rate for the moment estimations. Should be in range (0.0, 1.0). Default: ``(0.9, 0.999)``. eps (float, optional): term added to the denominator to improve numerical stability. Should be greater than 0. Default: ``1e-8``. weight_decay (float, optional): weight decay (L2 penalty). Default: ``0.``. amsgrad (bool, optional): whether to use the AMSGrad algorithm. Default: ``False``. Keyword Args: 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 `lr` is not int, float or Tensor. ValueError: If the `lr` is less than 0. ValueError: If the `eps` is less than 0.0. ValueError: If the `betas` is not in the range of [0, 1). ValueError: If the `weight_decay` is less than 0. Supported Platforms: ``Ascend`` Examples: >>> import mindspore >>> from mindspore import nn >>> from mindspore import mint >>> # Define the network structure of LeNet5. Refer to >>> # https://gitee.com/mindspore/docs/blob/master/docs/mindspore/code/lenet.py >>> net = LeNet5() >>> loss_fn = nn.SoftmaxCrossEntropyWithLogits(sparse=True) >>> optimizer = mint.optim.Adam(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-3, betas=(0.9, 0.999), eps=1e-8, weight_decay=0.0, amsgrad=False, *, maximize=False): _check_param_value(betas, eps, weight_decay, lr, amsgrad, maximize, self.cls_name) if lr < 0.0: raise ValueError("Invalid learning rate: {}".format(lr)) if eps < 0.0: raise ValueError("Invalid epsilon value: {}".format(eps)) if not 0.0 <= betas[0] < 1.0: raise ValueError("Invalid beta parameter at index 0: {}".format(betas[0])) if not 0.0 <= betas[1] < 1.0: raise ValueError("Invalid beta parameter at index 1: {}".format(betas[1])) if weight_decay < 0.0: raise ValueError("Invalid weight_decay value: {}".format(weight_decay)) defaults = dict(lr=lr, betas=betas, eps=eps, weight_decay=weight_decay, amsgrad=amsgrad, maximize=maximize) self.max_v_group = True super(Adam, self).__init__(params, defaults) self.exp_avg = self.parameters.clone(prefix="exp_avg", init='zeros') self.exp_avg_sq = self.parameters.clone(prefix="exp_avg_sq", init='zeros') self.state_step = Parameter(Tensor([0], mstype.float32), "state_step") self.increase_tensor = Tensor(1, mstype.float32) self.assignadd = P.AssignAdd() self.pow = P.Pow() def construct(self, gradients): self.assignadd(self.state_step, self.increase_tensor) for group_id, group in enumerate(self.param_groups): beta1, beta2 = group['betas'] maximize = group.get("maximize") start_id = self.group_start_id[group_id] end_id = self.group_start_id[group_id + 1] lr = group.get("lr") grads = tuple([grad if not maximize else mint.neg(grad) for grad in gradients[start_id: end_id]]) bias_correction1 = 1 - beta1 ** self.state_step bias_correction2 = 1 - beta2 ** self.state_step neg_step_size = -mint.div(lr, bias_correction1) sqrt_bias_correction2 = mint.sqrt(bias_correction2) grads = self._decay_weight(group.get("weight_decay"), self.parameters[start_id: end_id], grads) if group.get("amsgrad"): self.hyper_map(F.partial(_optim_adamw_opt, beta1, beta2, group.get("eps"), neg_step_size, sqrt_bias_correction2), self.parameters[start_id: end_id], grads, self.exp_avg[start_id: end_id], self.exp_avg_sq[start_id: end_id], group.get("max_exp_avg_sq")) else: self.hyper_map(F.partial(_optim_adamw_opt, beta1, beta2, group.get("eps"), neg_step_size, sqrt_bias_correction2), self.parameters[start_id: end_id], grads, self.exp_avg[start_id: end_id], self.exp_avg_sq[start_id: end_id]) return True