torch.nn.BCEWithLogitsLoss相当于sigmoid+torch.nn.BCELoss。代码示例如下，文章来源地址https://www.toymoban.com/news/detail-574398.html
import torch
import torch.nn as nn
BCEWithLogitsLoss = nn.BCEWithLogitsLoss()
BCELoss = nn.BCELoss()
x = torch.randn((1,))
y = torch.FloatTensor([1])
Loss_BCEWithLogits = BCEWithLogitsLoss(x, y)
Loss_BCE = BCELoss(torch.sigmoid(x), y)
print("BCEWithLogitsLoss:", Loss_BCEWithLogits)
print("BCELoss:", Loss_BCE)
"""
BCEWithLogitsLoss: tensor(0.2138)
BCELoss: tensor(0.2138)
"""文章来源：https://www.toymoban.com/news/detail-574398.html
到了这里，关于torch.nn.BCEWithLogitsLoss与torch.nn.BCELoss的文章就介绍完了。如果您还想了解更多内容，请在右上角搜索TOY模板网以前的文章或继续浏览下面的相关文章，希望大家以后多多支持TOY模板网！
