transpose() and t() in PyTorch

#pytorch #transpose #t #function

*Memos:

My post explains adjoint(), mH and mT.
My post explains reshape() and view().
My post explains permute().
My post explains movedim().

transpose() can get the 0D or more D transposed tensor of zero or more elements without losing data from the 0D or more D tensor of zero or more elements as shown below:

*Memos:

transpose() can be used with torch or a tensor.
The 1st argument(input) with torch or using a tensor(Required-Type:tensor of int, float, complex or bool).
The 2nd argument with torch or the 1st argument with a tensor is dim0(Required-Type:int).
The 3rd argument with torch or the 2nd argument with a tensor is dim1(Required-Type:int).
swapaxes() and swapdims() are the aliases of transpose().

import torch

my_tensor = torch.tensor([[[0, 1, 2], [3, 4, 5]],
                          [[6, 7, 8], [9, 10, 11]],
                          [[12, 13, 14], [15, 16, 17]],
                          [[18, 19, 20], [21, 22, 23]]])
torch.transpose(input=my_tensor, dim0=0, dim1=0)
my_tensor.transpose(dim0=0, dim1=0)
torch.transpose(input=my_tensor, dim0=1, dim1=1)
torch.transpose(input=my_tensor, dim0=2, dim1=2)
torch.transpose(input=my_tensor, dim0=1, dim1=-2)
torch.transpose(input=my_tensor, dim0=2, dim1=-1)
torch.transpose(input=my_tensor, dim0=2, dim1=-2)
torch.transpose(input=my_tensor, dim0=-1, dim1=2)
torch.transpose(input=my_tensor, dim0=-2, dim1=1)
torch.transpose(input=my_tensor, dim0=-1, dim1=-1)
torch.transpose(input=my_tensor, dim0=-2, dim1=-2)
# tensor([[[0, 1, 2], [3, 4, 5]],
#         [[6, 7, 8], [9, 10, 11]],
#         [[12, 13, 14], [15, 16, 17]],
#         [[18, 19, 20], [21, 22, 23]]])

torch.transpose(input=my_tensor, dim0=0, dim1=1)
torch.transpose(input=my_tensor, dim0=1, dim1=0)
torch.transpose(input=my_tensor, dim0=0, dim1=-2)
torch.transpose(input=my_tensor, dim0=-2, dim1=0)
# tensor([[[0, 1, 2], [6, 7, 8], [12, 13, 14], [18, 19, 20]],
#         [[3, 4, 5], [ 9, 10, 11], [15, 16, 17], [21, 22, 23]]])

torch.transpose(input=my_tensor, dim0=0, dim1=2)
torch.transpose(input=my_tensor, dim0=2, dim1=0)
torch.transpose(input=my_tensor, dim0=0, dim1=-1)
torch.transpose(input=my_tensor, dim0=-1, dim1=0)
# tensor([[[0, 6, 12, 18], [3, 9, 15, 21]],
#         [[1, 7, 13, 19], [4, 10, 16, 22]],
#         [[2, 8, 14, 20], [5, 11, 17, 23]]])

torch.transpose(input=my_tensor, dim0=1, dim1=2)
torch.transpose(input=my_tensor, dim0=2, dim1=1)
torch.transpose(input=my_tensor, dim0=1, dim1=-1)
torch.transpose(input=my_tensor, dim0=-1, dim1=1)
torch.transpose(input=my_tensor, dim0=-1, dim1=-2)
torch.transpose(input=my_tensor, dim0=-2, dim1=-1)
torch.transpose(input=my_tensor, dim0=-2, dim1=2)
# tensor([[[0, 3], [1, 4], [2, 5]],
#         [[6, 9], [7, 10], [8, 11]],
#         [[12, 15], [13, 16], [14, 17]],
#         [[18, 21], [19, 22], [20, 23]]])

my_tensor = torch.tensor([[[0., 1., 2.], [3., 4., 5.]],
                          [[6., 7., 8.], [9., 10., 11.]],
                          [[12., 13., 14.], [15., 16., 17.]],
                          [[18., 19., 20.], [21., 22., 23.]]])
torch.transpose(input=my_tensor, dim0=0, dim1=0)
# tensor([[[0., 1., 2.], [3., 4., 5.]],
#         [[6., 7., 8.], [9., 10., 11.]],
#         [[12., 13., 14.], [15., 16., 17.]],
#         [[18., 19., 20.], [21., 22., 23.]]])

my_tensor = torch.tensor([[[0.+0.j, 1.+0.j, 2.+0.j],
                           [3.+0.j, 4.+0.j, 5.+0.j]],
                          [[6.+0.j, 7.+0.j, 8.+0.j],
                           [9.+0.j, 10.+0.j, 11.+0.j]],
                          [[12.+0.j, 13.+0.j, 14.+0.j],
                           [15.+0.j, 16.+0.j, 17.+0.j]],
                          [[18.+0.j, 19.+0.j, 20.+0.j],
                           [21.+0.j, 22.+0.j, 23.+0.j]]])
torch.transpose(input=my_tensor, dim0=0, dim1=0)
# tensor([[[0.+0.j, 1.+0.j, 2.+0.j],
#          [3.+0.j, 4.+0.j, 5.+0.j]],
#         [[6.+0.j, 7.+0.j, 8.+0.j],
#          [9.+0.j, 10.+0.j, 11.+0.j]],
#         [[12.+0.j, 13.+0.j, 14.+0.j],
#          [15.+0.j, 16.+0.j, 17.+0.j]],
#         [[18.+0.j, 19.+0.j, 20.+0.j],
#          [21.+0.j, 22.+0.j, 23.+0.j]]])

my_tensor = torch.tensor([[[True, False, True], [True, False, True]],
                          [[False, True, False], [False, True, False]],
                          [[True, False, True], [True, False, True]],
                          [[False, True, False], [False, True, False]]])
torch.transpose(input=my_tensor, dim0=0, dim1=0)
# tensor([[[True, False, True], [True, False, True]],
#         [[False, True, False], [False, True, False]],
#         [[True, False, True], [True, False, True]],
#         [[False, True, False], [False, True, False]]])

t() can get the 0D, 1D or 2D transposed tensor of zero or more elements without losing data from the 0D, 1D or 2D tensor of zero or more elements as shown below:

*Memos:

t() can be used with torch or a tensor.
The 1st argument(input) with torch or using a tensor(Required-Type:tensor of int, float, complex or bool).
t() is equivalent to transpose(input=my_tensor, dim0=0, dim1=1).

import torch

my_tensor = torch.tensor(0)

torch.t(input=my_tensor)
my_tensor.t()
# tensor(0)

my_tensor = torch.tensor([0, 1, 2])

torch.t(input=my_tensor)
# tensor([0, 1, 2])

my_tensor = torch.tensor([[0, 1, 2], [3, 4, 5],
                          [6, 7, 8], [9, 10, 11]])
torch.t(input=my_tensor)
# tensor([[0, 3, 6, 9],
#         [1, 4, 7, 10],
#         [2, 5, 8, 11]])

my_tensor = torch.tensor([[0., 1., 2.], [3., 4., 5.],
                          [6., 7., 8.], [9., 10., 11.]])
torch.t(input=my_tensor)
# tensor([[0., 3., 6., 9.],
#         [1., 4., 7., 10.],
#         [2., 5., 8., 11.]])

my_tensor = torch.tensor([[0.+0.j, 1.+0.j, 2.+0.j],
                          [3.+0.j, 4.+0.j, 5.+0.j],
                          [6.+0.j, 7.+0.j, 8.+0.j],
                          [9.+0.j, 10.+0.j, 11.+0.j]])
torch.t(input=my_tensor)
# tensor([[0.+0.j, 3.+0.j, 6.+0.j, 9.+0.j],
#         [1.+0.j, 4.+0.j, 7.+0.j, 10.+0.j],
#         [2.+0.j, 5.+0.j, 8.+0.j, 11.+0.j]])

my_tensor = torch.tensor([[True, False, True], [True, False, True],
                          [False, True, False], [False, True, False]])
torch.t(input=my_tensor)
# tensor([[True, True, False, False],
#         [False, False, True, True],
#         [True, True, False, False]])

DEV Community

transpose() and t() in PyTorch

Top comments (0)

Read next

Migrate MongoDB to MySQL using Prism Key Differences

Automate and Elevate: Integrating GitHub Actions with SonarCloud for Superior Code Quality

Can you develop and test AR applications using the Unity app for iPhone?

Create your own 📷 image compressor & to webp convertor using HTML, CSS, JavaScript from scratch 🚀