[resolved] Difference b/w Long Tensor and Float Tensor?

1

I was going through the transfer learning tutorial , and I noticed this
_ , prediction = torch.max(outputs.data , 1)

So I wondered what the first return data was.
I ran the following code , can someone help me understand the output difference

a = Variable(tt.Tensor([[1,2,.01],
    ...:                        [5,0.1,2],
    ...:                        [4,5,6]]))

In [16]: a
Out[16]: 
Variable containing:
 1.0000  2.0000  0.0100
 5.0000  0.1000  2.0000
 4.0000  5.0000  6.0000
[torch.FloatTensor of size 3x3]

  p = torch.max(a.data , 1)
print(p)
(
 2
 5
 6
[torch.FloatTensor of size 3x1]
, 
 1
 0
 2
[torch.LongTensor of size 3x1]
)

I understood the first output , but can not understand the second result ( LongTensor)

2

Thanks , Got the answer after searching the docs !

1 Like