BCEWithLogitsLoss

最新推荐文章于 2024-01-12 13:56:51 发布

*Major*

最新推荐文章于 2024-01-12 13:56:51 发布

阅读量688

点赞数

本文链接：https://blog.csdn.net/qq_41375318/article/details/112184901

版权

$B C E W i t h L o g i t s L o s s$

一 Pytorch实现

torch.nn.BCEWithLogitsLoss(
weight: Optional[torch.Tensor] = None, 
size_average=None, 
reduce=None, 
reduction: str = 'mean', 
pos_weight: Optional[torch.Tensor] = None)

在这里插入图片描述

target = torch.ones([10, 64], dtype=torch.float32)  # 64 classes, batch size = 10
output = torch.full([10, 64], 1.5)  # A prediction (logit)
pos_weight = torch.ones([64])  # All weights are equal to 1
criterion = torch.nn.BCEWithLogitsLoss(pos_weight=pos_weight)
criterion(output, target)  # -log(sigmoid(1.5))

在这里插入图片描述

import torch
from torch import nn
target =  torch.tensor([[1,2,3]],dtype=torch.float)
output =  torch.tensor([[1,2,30]],dtype=torch.float)
criterion = torch.nn.BCEWithLogitsLoss()
criterion(output, target)  # -log(sigmoid(1.5))