ECC-WBCE 损失函数代码

ECC-WBCE（Enhanced Cross-Entropy with Weighted Binary Cross-Entropy）是一种用于二分类任务的损失函数，它结合了交叉熵损失和加权二元交叉熵损失。下面是ECC-WBCE损失函数的代码示例：

import torch
import torch.nn as nn

class ECC_WBCE_Loss(nn.Module):
    def __init__(self, alpha=0.5, beta=0.5, epsilon=1e-7):
        super(ECC_WBCE_Loss, self).__init__()
        self.alpha = alpha
        self.beta = beta
        self.epsilon = epsilon

    def forward(self, y_pred, y_true):
        y_pred = torch.clamp(y_pred, self.epsilon, 1.0 - self.epsilon)
        loss_ce = - (self.alpha * y_true * torch.log(y_pred) + (1 - self.alpha) * (1 - y_true) * torch.log(1 - y_pred))
        loss_wbce = - (self.beta * y_true * torch.log(y_pred) + (1 - self.beta) * (1 - y_true) * torch.log(1 - y_pred))
        loss = loss_ce + loss_wbce
        return loss.mean()

在上述代码中，`ECC_WBCE_Loss`是一个继承自`nn.Module`的自定义损失函数类。在`forward`方法中，首先对预测值进行限制，避免取对数时出现无穷大或无穷小的情况。然后分别计算交叉熵损失和加权二元交叉熵损失，并将两者相加得到最终的损失值。