文字版说明:

Variable在 Pytorch0.4.0之后以及被tensor(requires_grad=True)所取代了.

The Variable API has been deprecated in Pytorch0.4.0: Variables are no longer necessary to use autograd with tensors. Autograd automatically supports Tensors with requires_grad set to True. Below please find a quick guide on what has changed:

• Variable(tensor) and Variable(tensor, requires_grad) still work as expected, but they return Tensors instead of Variables.
• var.data is the same thing as tensor.data.
• Methods such as var.backward(), var.detach(), var.register_hook() now work on tensors with the same method names.

In addition, one can now create tensors with requires_grad=True using factory methods such as torch.randn(), torch.zeros(), torch.ones(), and others like the following:

autograd_tensor = torch.randn((2, 3, 4), requires_grad=True)

You can also find some explanations about the difference between a tensor and a Variable in What is the difference between Tensors and Variables in Pytorch?

功能相同代码验证:(验证可以同样来使用)

Variable版本:

requires_grad=True版本: