Pytorch中的Variable

注：转化为此变量才能进行反向传播，tensor本身不能进行反向传播
Variable可以把输出的Tensor变成一个输入变量，这样梯度就不会回传了。
推荐阅读

import torch

import numpy as np

from torch.autograd import Variable

tensor = torch.FloatTensor([[1, 2],
                            [3, 4]])
variable = Variable(tensor, requires_grad=True)

# tensor cann't backward but varible may
t_out = torch.mean(tensor * tensor)
v_out = torch.mean(variable * variable)
print(t_out)
print(v_out)

v_out.backward()

# grad values
print(variable.grad)

# variable style is  variable but variable.data style is tensor
print(variable.data)

# tensor to numpy
print(variable.data.numpy())