PyTorch 中的标签平滑

新手上路,请多包涵

我正在使用迁移学习为 斯坦福汽车 数据集构建一个 ResNet-18 分类模型。我想实施 标签平滑 来惩罚过度自信的预测并提高泛化能力。

TensorFlowCrossEntropyLoss 中有一个简单的关键字参数。有没有人为 PyTorch 构建了一个我可以即插即用的类似功能?

原文由 Jared Nielsen 发布,翻译遵循 CC BY-SA 4.0 许可协议

阅读 731
2 个回答

多类神经网络的泛化和学习速度通常可以通过使用 软目标 来显着提高,软目标是 硬目标加权平均值 和标签上的 均匀分布。以这种方式平滑标签可防止网络变得过于自信,标签平滑已用于许多最先进的模型,包括图像分类、语言翻译和语音识别。


标签平滑 已经在 Tensorflow 的交叉熵损失函数中实现。 二元交叉熵分类 交叉熵。但是目前,在 PyTorch 中没有正式实现 Label Smoothing 。然而,目前正在进行积极的讨论,希望它能提供一个官方包。这是讨论主题: 问题#7455

在这里,我们将从 PyTorch 实践者那里带来一些可用的 标签平滑 (LS) 的最佳实现。基本上,有很多方法可以实现 LS 。请参考这个关于这个的具体讨论,一个在 这里另一个在这里。在这里,我们将以 两种 独特的方式实现,每种方式都有两个版本;所以总共 4

选项 1:CrossEntropyLossWithProbs

这样,它接受了 one-hot 目标向量。用户必须手动平滑他们的目标矢量。它可以在 with torch.no_grad() 范围内完成,因为它暂时将所有 requires_grad 标志设置为 false。

  1. 杨德文资料来源
import torch
import numpy as np
import torch.nn as nn
import torch.nn.functional as F
from torch.autograd import Variable
from torch.nn.modules.loss import _WeightedLoss

class LabelSmoothingLoss(nn.Module):
    def __init__(self, classes, smoothing=0.0, dim=-1, weight = None):
        """if smoothing == 0, it's one-hot method
           if 0 < smoothing < 1, it's smooth method
        """
        super(LabelSmoothingLoss, self).__init__()
        self.confidence = 1.0 - smoothing
        self.smoothing = smoothing
        self.weight = weight
        self.cls = classes
        self.dim = dim

    def forward(self, pred, target):
        assert 0 <= self.smoothing < 1
        pred = pred.log_softmax(dim=self.dim)

        if self.weight is not None:
            pred = pred * self.weight.unsqueeze(0)

        with torch.no_grad():
            true_dist = torch.zeros_like(pred)
            true_dist.fill_(self.smoothing / (self.cls - 1))
            true_dist.scatter_(1, target.data.unsqueeze(1), self.confidence)
        return torch.mean(torch.sum(-true_dist * pred, dim=self.dim))

此外,我们在 self. smoothing 上添加了一个断言复选标记,并为此实现添加了损失加权支持。

  1. Shital Shah : 资料来源

Shital 已经在这里发布了答案。这里我们要指出的是,这个实现类似于 Devin Yang 的上述实现。然而,在这里我们提到他的代码,并最小化了一点 code syntax

 class SmoothCrossEntropyLoss(_WeightedLoss):
    def __init__(self, weight=None, reduction='mean', smoothing=0.0):
        super().__init__(weight=weight, reduction=reduction)
        self.smoothing = smoothing
        self.weight = weight
        self.reduction = reduction

    def k_one_hot(self, targets:torch.Tensor, n_classes:int, smoothing=0.0):
        with torch.no_grad():
            targets = torch.empty(size=(targets.size(0), n_classes),
                                  device=targets.device) \
                                  .fill_(smoothing /(n_classes-1)) \
                                  .scatter_(1, targets.data.unsqueeze(1), 1.-smoothing)
        return targets

    def reduce_loss(self, loss):
        return loss.mean() if self.reduction == 'mean' else loss.sum() \
        if self.reduction == 'sum' else loss

    def forward(self, inputs, targets):
        assert 0 <= self.smoothing < 1

        targets = self.k_one_hot(targets, inputs.size(-1), self.smoothing)
        log_preds = F.log_softmax(inputs, -1)

        if self.weight is not None:
            log_preds = log_preds * self.weight.unsqueeze(0)

        return self.reduce_loss(-(targets * log_preds).sum(dim=-1))

查看

import torch
import numpy as np
import torch.nn as nn
import torch.nn.functional as F
from torch.autograd import Variable
from torch.nn.modules.loss import _WeightedLoss

if __name__=="__main__":
    # 1. Devin Yang
    crit = LabelSmoothingLoss(classes=5, smoothing=0.5)
    predict = torch.FloatTensor([[0, 0.2, 0.7, 0.1, 0],
                                 [0, 0.9, 0.2, 0.2, 1],
                                 [1, 0.2, 0.7, 0.9, 1]])
    v = crit(Variable(predict),
             Variable(torch.LongTensor([2, 1, 0])))
    print(v)

    # 2. Shital Shah
    crit = SmoothCrossEntropyLoss(smoothing=0.5)
    predict = torch.FloatTensor([[0, 0.2, 0.7, 0.1, 0],
                                 [0, 0.9, 0.2, 0.2, 1],
                                 [1, 0.2, 0.7, 0.9, 1]])
    v = crit(Variable(predict),
             Variable(torch.LongTensor([2, 1, 0])))
    print(v)

tensor(1.4178)
tensor(1.4178)


选项 2:LabelSmoothingCrossEntropyLoss

通过这种方式,它接受目标向量并使用不手动平滑目标向量,而是内置模块负责标签平滑。它允许我们根据 F.nll_loss 实施标签平滑。

(一种)。 Wangleiofficial : 来源- (AFAIK), Original Poster

(二). Datasaurus : Source - 添加了加权支持

此外,我们略微减少了代码编写,使其更加简洁。

 class LabelSmoothingLoss(torch.nn.Module):
    def __init__(self, smoothing: float = 0.1,
                 reduction="mean", weight=None):
        super(LabelSmoothingLoss, self).__init__()
        self.smoothing   = smoothing
        self.reduction = reduction
        self.weight    = weight

    def reduce_loss(self, loss):
        return loss.mean() if self.reduction == 'mean' else loss.sum() \
         if self.reduction == 'sum' else loss

    def linear_combination(self, x, y):
        return self.smoothing * x + (1 - self.smoothing) * y

    def forward(self, preds, target):
        assert 0 <= self.smoothing < 1

        if self.weight is not None:
            self.weight = self.weight.to(preds.device)

        n = preds.size(-1)
        log_preds = F.log_softmax(preds, dim=-1)
        loss = self.reduce_loss(-log_preds.sum(dim=-1))
        nll = F.nll_loss(
            log_preds, target, reduction=self.reduction, weight=self.weight
        )
        return self.linear_combination(loss / n, nll)

  1. NVIDIA/深度学习示例来源
class LabelSmoothing(nn.Module):
    """NLL loss with label smoothing.
    """
    def __init__(self, smoothing=0.0):
        """Constructor for the LabelSmoothing module.
        :param smoothing: label smoothing factor
        """
        super(LabelSmoothing, self).__init__()
        self.confidence = 1.0 - smoothing
        self.smoothing = smoothing

    def forward(self, x, target):
        logprobs = torch.nn.functional.log_softmax(x, dim=-1)
        nll_loss = -logprobs.gather(dim=-1, index=target.unsqueeze(1))
        nll_loss = nll_loss.squeeze(1)
        smooth_loss = -logprobs.mean(dim=-1)
        loss = self.confidence * nll_loss + self.smoothing * smooth_loss
        return loss.mean()

查看

if __name__=="__main__":
    # Wangleiofficial
    crit = LabelSmoothingLoss(smoothing=0.3, reduction="mean")
    predict = torch.FloatTensor([[0, 0.2, 0.7, 0.1, 0],
                                 [0, 0.9, 0.2, 0.2, 1],
                                 [1, 0.2, 0.7, 0.9, 1]])

    v = crit(Variable(predict),
             Variable(torch.LongTensor([2, 1, 0])))
    print(v)

    # NVIDIA
    crit = LabelSmoothing(smoothing=0.3)
    predict = torch.FloatTensor([[0, 0.2, 0.7, 0.1, 0],
                                 [0, 0.9, 0.2, 0.2, 1],
                                 [1, 0.2, 0.7, 0.9, 1]])
    v = crit(Variable(predict),
             Variable(torch.LongTensor([2, 1, 0])))
    print(v)

tensor(1.3883)
tensor(1.3883)


更新: 正式添加

torch.nn.CrossEntropyLoss(weight=None, size_average=None,
                          ignore_index=- 100, reduce=None,
                          reduction='mean', label_smoothing=0.0)

原文由 Innat 发布,翻译遵循 CC BY-SA 4.0 许可协议

我一直在寻找源自 _Loss 的选项,就像 PyTorch 中的其他损失类一样,并尊重基本参数,例如 reduction 。不幸的是,我找不到直接的替代品,所以最终自己写了。但是,我还没有对此进行全面测试:

 import torch
from torch.nn.modules.loss import _WeightedLoss
import torch.nn.functional as F

class SmoothCrossEntropyLoss(_WeightedLoss):
    def __init__(self, weight=None, reduction='mean', smoothing=0.0):
        super().__init__(weight=weight, reduction=reduction)
        self.smoothing = smoothing
        self.weight = weight
        self.reduction = reduction

    @staticmethod
    def _smooth_one_hot(targets:torch.Tensor, n_classes:int, smoothing=0.0):
        assert 0 <= smoothing < 1
        with torch.no_grad():
            targets = torch.empty(size=(targets.size(0), n_classes),
                    device=targets.device) \
                .fill_(smoothing /(n_classes-1)) \
                .scatter_(1, targets.data.unsqueeze(1), 1.-smoothing)
        return targets

    def forward(self, inputs, targets):
        targets = SmoothCrossEntropyLoss._smooth_one_hot(targets, inputs.size(-1),
            self.smoothing)
        lsm = F.log_softmax(inputs, -1)

        if self.weight is not None:
            lsm = lsm * self.weight.unsqueeze(0)

        loss = -(targets * lsm).sum(-1)

        if  self.reduction == 'sum':
            loss = loss.sum()
        elif  self.reduction == 'mean':
            loss = loss.mean()

        return loss

其他选项:

原文由 Shital Shah 发布,翻译遵循 CC BY-SA 4.0 许可协议

撰写回答
你尚未登录,登录后可以
  • 和开发者交流问题的细节
  • 关注并接收问题和回答的更新提醒
  • 参与内容的编辑和改进,让解决方法与时俱进
推荐问题