渣渣屋

文章目录标量函数backward向量函数backeward，及gradient参数解析 标量函数backward import torch from torch.autograd import Variable import torch.nn as nn import torch.nn.functional as F #反向传播 x = torch.ones(2, 2, requires_grad=True) y = x + 2 z = y * y out = z.mean() # 如果没有下面这一行，