Vanilla SGD

最新推荐文章于 2023-04-05 16:06:35 发布

AKA_3

最新推荐文章于 2023-04-05 16:06:35 发布

阅读量988

点赞数

分类专栏： SGD 文章标签：算法

本文链接：https://blog.csdn.net/weixin_44514137/article/details/117788647

版权

SGD 专栏收录该内容

1 篇文章

订阅专栏

梯度下降

梯度下降是目前神经网络中使用最为广泛的优化算法之一。

利用梯度下降求解的时候遵循这样一个模式，梯度下降是指，在给定待优化的模型参数 $\theta\in \mathbb{R}^{d}$ 和目标函数 $J(\theta )$ 后，算法通过沿梯度 $\triangledown_{\theta} J(\theta)$ 的相反方向更新 $\theta$ 来最小化 $J(\theta )$ 。学习率 $\eta$ 决定了每一时刻的更新步长。对于每一个时刻 $t$ ，我们可以用下述步骤描述梯度下降的流程:
（1）计算损失函数关于参数 𝜽 的梯度：

$g_{t}=\triangledown_{\theta} J(\theta)$

（2）根据历史梯度计算一阶动量和二阶动量：

$m_{t}=\phi (g_{1},g_{2},\cdot \cdot \cdot ,g_{t})$

$v_{t}=\psi (g_{1},g_{2},\cdot \cdot \cdot ,g_{t})$

PS:本文所提到的算法朴素SGD并未引入动量的概念

（3）计算参数更新量，其中 $\omega$ 为防止分母为0的平滑项，通常取 $1\times 10^{-8}$ ：

$\Delta \theta_{t}=\eta \cdot \frac{m_{t}}{\sqrt{V_{t}+\omega }}$

（4）进行参数更新：

$\theta_{t+1}=\theta_{t}-\eta \cdot \frac{m_{t}}{\sqrt{V_{t}+\omega }}$

朴素随机梯度下降（Vanilla SGD）

朴素SGD没有动量的概念，即 $m_{t}=g_{t}$ ， $V_{t} = I^{2}$ ， $\omega =0$ ,原公式简化为：

$\theta_{t+1}=\theta_{t}-\eta \cdot g_{t}$

通过将当前权重减去其梯度的因子（即 $\eta$ ，学习率）来更新权重。

SGD 的缺点在于收敛速度慢，在下降的过程中会出现震荡，特别是容易陷入局部最优点或者是鞍点。并且，如何合理的选择学习率是 SGD 的一大难点。

Vanilla SGD的PyTorch方法

class SGD(Optimizer):
    r"""Implements stochastic gradient descent (optionally with momentum).

    Nesterov momentum is based on the formula from
    `On the importance of initialization and momentum in deep learning`__.

    Args:
        params (iterable): iterable of parameters to optimize or dicts defining
            parameter groups
        lr (float): learning rate
        momentum (float, optional): momentum factor (default: 0)
        weight_decay (float, optional): weight decay (L2 penalty) (default: 0)
        dampening (float, optional): dampening for momentum (default: 0)
        nesterov (bool, optional): enables Nesterov momentum (default: False)

    Example:
        >>> optimizer = torch.optim.SGD(model.parameters(), lr=0.1, momentum=0.9)
        >>> optimizer.zero_grad()
        >>> loss_fn(model(input), target).backward()
        >>> optimizer.step()

    __ http://www.cs.toronto.edu/%7Ehinton/absps/momentum.pdf

    .. note::
        The implementation of SGD with Momentum/Nesterov subtly differs from
        Sutskever et. al. and implementations in some other frameworks.

        Considering the specific case of Momentum, the update can be written as

        .. math::
            \begin{aligned}
                v_{t+1} & = \mu * v_{t} + g_{t+1}, \\
                p_{t+1} & = p_{t} - \text{lr} * v_{t+1},
            \end{aligned}

        where :math:`p`, :math:`g`, :math:`v` and :math:`\mu` denote the 
        parameters, gradient, velocity, and momentum respectively.

        This is in contrast to Sutskever et. al. and
        other frameworks which employ an update of the form

        .. math::
            \begin{aligned}
                v_{t+1} & = \mu * v_{t} + \text{lr} * g_{t+1}, \\
                p_{t+1} & = p_{t} - v_{t+1}.
            \end{aligned}

        The Nesterov version is analogously modified.
    """

    def __init__(self, params, lr=required, momentum=0, dampening=0,
                 weight_decay=0, nesterov=False):
        if lr is not required and lr < 0.0:
            raise ValueError("Invalid learning rate: {}".format(lr))
        if momentum < 0.0:
            raise ValueError("Invalid momentum value: {}".format(momentum))
        if weight_decay < 0.0:
            raise ValueError("Invalid weight_decay value: {}".format(weight_decay))

        defaults = dict(lr=lr, momentum=momentum, dampening=dampening,
                        weight_decay=weight_decay, nesterov=nesterov)
        if nesterov and (momentum <= 0 or dampening != 0):
            raise ValueError("Nesterov momentum requires a momentum and zero dampening")
        super(SGD, self).__init__(params, defaults)

    def __setstate__(self, state):
        super(SGD, self).__setstate__(state)
        for group in self.param_groups:
            group.setdefault('nesterov', False)

    @torch.no_grad()
    def step(self, closure=None):
        """Performs a single optimization step.

        Args:
            closure (callable, optional): A closure that reevaluates the model
                and returns the loss.
        """
        loss = None
        if closure is not None:
            with torch.enable_grad():
                loss = closure()

        for group in self.param_groups:
            params_with_grad = []
            d_p_list = []
            momentum_buffer_list = []
            weight_decay = group['weight_decay']
            momentum = group['momentum']
            dampening = group['dampening']
            nesterov = group['nesterov']
            lr = group['lr']

            for p in group['params']:
                if p.grad is not None:
                    params_with_grad.append(p)
                    d_p_list.append(p.grad)

                    state = self.state[p]
                    if 'momentum_buffer' not in state:
                        momentum_buffer_list.append(None)
                    else:
                        momentum_buffer_list.append(state['momentum_buffer'])

            F.sgd(params_with_grad,
                  d_p_list,
                  momentum_buffer_list,
                  weight_decay,
                  momentum,
                  lr,
                  dampening,
                  nesterov)

            # update momentum_buffers in state
            for p, momentum_buffer in zip(params_with_grad, momentum_buffer_list):
                state = self.state[p]
                state['momentum_buffer'] = momentum_buffer

        return loss