transpose and t in PyTorch

#python #pytorch #transpose #t

*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

Ollama - Custom Model - llama3.2

Django Rest framework with Swagger

#? List vs Tuples in python

DUCK (file structure) YOU!