PyTorch基本用法(二)——Variable

客：noahsnail.com  |  CSDN  |  简书

本文主要是PyTorch中Variable变量的一些用法。

import torchfrom torch.autograd import Variabletensor = torch.FloatTensor([[1, 2], [3, 4]])# 定义Variable, requires_grad用来指定是否需要计算梯度variable = Variable(tensor, requires_grad = True)print tensorprint variable

 1  2 3  4[torch.FloatTensor of size 2x2]Variable containing: 1  2 3  4[torch.FloatTensor of size 2x2]

# 计算x^2的均值tensor_mean = torch.mean(tensor * tensor)variable_mean = torch.mean(variable * variable)print tensor_meanprint variable_mean

7.5Variable containing: 7.5000[torch.FloatTensor of size 1]

# variable进行反向传播# 梯度计算如下：# variable_mean = 1/4 * sum(variable * variable)# d(variable_mean)/d(variable) = 1/4 * 2 * variable = 1/2 * variablevariable_mean.backward()# 输出variable中的梯度print variable.grad

Variable containing: 0.5000  1.0000 1.5000  2.0000[torch.FloatTensor of size 2x2]

# *表示逐元素点乘,不是矩阵乘法print tensor * tensorprint variable * variable

  1   4  9  16[torch.FloatTensor of size 2x2]Variable containing:  1   4  9  16[torch.FloatTensor of size 2x2]

# 输出variable中的data, data是tensorprint variable.data

 1  2 3  4[torch.FloatTensor of size 2x2]