pytorch

损失函数

1. CrossEntropyloss

a.交叉熵损失函数，常用于分类
b.用这个loss前面不需要加 softmax层
c.该函数限制了target的类型为torch.LongTensor

import torch as t
from torch import nn
from torch.autograd import Variable as V
# batch_size=4, 计算每个类别分数（二分类）
output = V(t.randn(4,2)) # batch_size * C=(batch_size, C)
# target必须是LongTensor!
target =V(t.Tensor([1,0,1,1])).long()

criterion = nn.CrossEntropyLoss()
loss = criterion(output, target)
print('loss', loss)

output: loss tensor(1.0643)

2. toch.nn.MSELoss

均方损失函数，类似于nn.L1Loss函数：

import torch
loss_fn = torch.nn.MSELoss(reduce=False, size_average=False)
input = torch.autograd.Variable(torch.randn(3,4))
target = torch.autograd.Variable(torch.randn(3,4))
loss = loss_fn(input, target)
print(input); print(target); print(loss)
print(input.size(), target.size(), loss.size())

output: