How does transpose work in Pytorch?

1

I was doing a transpose of tensors of rank 3 and according to transpose rule for rank 2 tensors which follow simple matrix transpose rule Aij = Aji. But when i transposed a rank 3 tensor i ended up with a different output given below.
Can someone explain to me how it is working under the hood?

a = torch.tensor(
        [[[1., 1., 1.],
         [1., 1., 1.],
         [1., 1., 1.]],

        [[0., 0., 0.],
         [0., 0., 0.],
         [0., 0., 0.]],

        [[1., 1., 1.],
         [1., 1., 1.],
         [1., 1., 1.]]])

Printing the tensor and the shape gives the following:

print(a)
print(a.shape)
tensor([[[1., 1., 1.],
         [1., 1., 1.],
         [1., 1., 1.]],

        [[0., 0., 0.],
         [0., 0., 0.],
         [0., 0., 0.]],

        [[1., 1., 1.],
         [1., 1., 1.],
         [1., 1., 1.]]])
torch.Size([3, 3, 3])

Transposing with a.T we get

tensor([[[1., 0., 1.],
         [1., 0., 1.],
         [1., 0., 1.]],

        [[1., 0., 1.],
         [1., 0., 1.],
         [1., 0., 1.]],

        [[1., 0., 1.],
         [1., 0., 1.],
         [1., 0., 1.]]])