pytorch torch.nn.L1Loss

应用

# 1)计算绝对差总和：|0-1|+|1-1|+|2-1|+|3-1|=4；
# 2)再平均：4/4=1。
import torch
import torch.nn as nn
sample = torch.tensor([[1.,1.],[1.,1.]])
target = torch.tensor([[0.,1.],[2.,3.]])
criterion = nn.L1Loss()
loss = criterion(sample, target)

API

CLASS torch.nn.L1Loss(size_average=None, reduce=None, reduction: str = 'mean')

mean absolute error (MAE)

参数	描述
size_average (bool, optional)
reduce (bool, optional)
reduction (string, optional)	‘none’ ， ‘mean’ ， ‘sum’，‘none’

from torch import nn
import torch
loss = nn.L1Loss()
input = torch.randn(3, 5, requires_grad=True)
target = torch.randn(3, 5)
output = loss(input, target)
output.backward()