pytorch的nn.MSELoss损失函数

MSE是mean squared error的缩写，即平均平方误差，简称均方误差。

MSE是逐元素计算的，计算公式为：

旧版的nn.MSELoss()函数有reduce、size_average两个参数，新版的只有一个reduction参数了，功能是一样的。reduction的意思是维度要不要缩减，以及怎么缩减，有三个选项：

• 'none': no reduction will be applied.
• 'mean': the sum of the output will be divided by the number of elements in the output.
• 'sum': the output will be summed.

如果不设置reduction参数，默认是'mean'。

程序示例： 

?

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

import torch import torch.nn as nn   a = torch.tensor([[1, 2], [3, 4]], dtype=torch.float32) b = torch.tensor([[3, 5], [8, 6]], dtype=torch.float32)   loss_fn1 = torch.nn.MSELoss(reduction='none') loss1 = loss_fn1(a, b) print(loss1)   # 输出结果：tensor([[ 4.,  9.],                #                 [25.,  4.]])   loss_fn2 = torch.nn.MSELoss(reduction='sum') loss2 = loss_fn2(a, b) print(loss2)   # 输出结果：tensor(42.)     loss_fn3 = torch.nn.MSELoss(reduction='mean') loss3 = loss_fn3(a, b) print(loss3)   # 输出结果：tensor(10.5000)

 

对于三维的输入也是一样的：

?

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

a = torch.randint(0, 9, (2, 2, 3)).float() b = torch.randint(0, 9, (2, 2, 3)).float() print('a:\n', a) print('b:\n', b)   loss_fn1 = torch.nn.MSELoss(reduction='none') loss1 = loss_fn1(a, b) print('loss_none:\n', loss1)   loss_fn2 = torch.nn.MSELoss(reduction='sum') loss2 = loss_fn2(a, b) print('loss_sum:\n', loss2)     loss_fn3 = torch.nn.MSELoss(reduction='mean') loss3 = loss_fn3(a, b) print('loss_mean:\n', loss3)

运行结果：

 

 参考资料：

pytorch的nn.MSELoss损失函数