torch.stack()作用,ValueError:only one element tensors can be converted to Python scalars

把list转换成pytorch的tensor对象时，出现以下错误
ValueError:only one element tensors can be converted to Python scalars

具体如下，torch.as_tensor()和torch.tensor()都转不了。

>>> t1 = torch.tensor([0., 1.])
>>> t2 = torch.tensor([2., 3.])
>>> tensor_list = [t1, t2]
>>> torch.as_tensor(tensor_list)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ValueError: only one element tensors can be converted to Python scalars
>>> torch.tensor(tensor_list)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ValueError: only one element tensors can be converted to Python scalars

   
   

   
   
1
2
3
4
5
6
7
8
9
10
11

准确地说，torch.tensor()不能把包含tensor的list转成tensor，纯list就可以转tensor，不信看下面代码，我也不知道是为什么。

>>> t1 = [0., 1.]
>>> t2 = [2., 3.]
>>> tensor_list = [t1, t2]
>>> torch.tensor(tensor_list)
tensor([[0., 1.],
        [2., 3.]])

   
   

   
   
1
2
3
4
5
6

解决办法

torch.stack()

>>> t1 = torch.tensor([0., 1.])
>>> t2 = torch.tensor([2., 3.])
>>> tensor_list = [t1, t2]
>>> torch.stack(tensor_list)
tensor([[0., 1.],
        [2., 3.]])

   
   

   
   
1
2
3
4
5
6

同类错误
如果list里面的tensor是int类型的，用torch.as_tensor()和torch.tensor()转tensor会报以下错误，同样用torch.stack()可以解决。
TypeError: only integer tensors of a single element can be converted to an index