DEV Community

Super Kai (Kazuya Ito)
Super Kai (Kazuya Ito)

Posted on • Edited on

1

flip in PyTorch

Buy Me a Coffee

*My post explains flipud().

flip() can get the 0D or more D tensor of reversed zero or more elements from the 0D or more D tensor of zero or more elements as shown below:

*Memos:

  • flip() 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 or more arguments with a tensor are dims(Required-Type:int, tuple of int or list of int). *Each number must be unique.
import torch

my_tensor = torch.tensor(2) # 0D tensor

torch.flip(input=my_tensor, dims=(0,))
my_tensor.flip(dims=(0,))
my_tensor.flip(0)
torch.flip(input=my_tensor, dims=(-1,))
# tensor(2)

my_tensor = torch.tensor([2, 7, 4]) # 1D tensor

torch.flip(input=my_tensor, dims=(0,))
torch.flip(input=my_tensor, dims=(-1,))
# tensor([4, 7, 2])

my_tensor = torch.tensor([[2, 7, 4], [8, 3, 2]]) # 2D tensor

torch.flip(input=my_tensor, dims=(0,))
torch.flip(input=my_tensor, dims=(-2,))
# tensor([[8, 3, 2], [2, 7, 4]])

torch.flip(input=my_tensor, dims=(1,))
torch.flip(input=my_tensor, dims=(-1,))
# tensor([[4, 7, 2], [2, 3, 8]])

torch.flip(input=my_tensor, dims=(0, 1))
torch.flip(input=my_tensor, dims=(0, -1))
torch.flip(input=my_tensor, dims=(1, 0))
torch.flip(input=my_tensor, dims=(1, -2))
torch.flip(input=my_tensor, dims=(-1, 0))
torch.flip(input=my_tensor, dims=(-1, -2))
torch.flip(input=my_tensor, dims=(-2, 1))
torch.flip(input=my_tensor, dims=(-2, -1))
# tensor([[2, 3, 8], [4, 7, 2]])

my_tensor = torch.tensor([[[2, 7, 4], [8, 3, 2]], # 3D tensor
                          [[5, 0, 8], [3, 6, 1]]])
torch.flip(input=my_tensor, dims=(0,))
torch.flip(input=my_tensor, dims=(-3,))
# tensor([[[5, 0, 8], [3, 6, 1]],
#         [[2, 7, 4], [8, 3, 2]]])

torch.flip(input=my_tensor, dims=(1,))
torch.flip(input=my_tensor, dims=(-2,))
# tensor([[[8, 3, 2], [2, 7, 4]],
#         [[3, 6, 1], [5, 0, 8]]])

torch.flip(input=my_tensor, dims=(2,))
torch.flip(input=my_tensor, dims=(-1,))
# tensor([[[4, 7, 2], [2, 3, 8]],
#         [[8, 0, 5], [1, 6, 3]]])

torch.flip(input=my_tensor, dims=(0, 1))
torch.flip(input=my_tensor, dims=(0, -2))
torch.flip(input=my_tensor, dims=(1, 0))
torch.flip(input=my_tensor, dims=(1, -3))
torch.flip(input=my_tensor, dims=(-2, 0))
torch.flip(input=my_tensor, dims=(-2, -3))
torch.flip(input=my_tensor, dims=(-3, 1))
torch.flip(input=my_tensor, dims=(-3, -2))
# tensor([[[3, 6, 1], [5, 0, 8]],
#         [[8, 3, 2], [2, 7, 4]]])

torch.flip(input=my_tensor, dims=(0, 2))
torch.flip(input=my_tensor, dims=(0, -1))
torch.flip(input=my_tensor, dims=(2, 0))
torch.flip(input=my_tensor, dims=(2, -3))
torch.flip(input=my_tensor, dims=(-1, 0))
torch.flip(input=my_tensor, dims=(-1, -3))
torch.flip(input=my_tensor, dims=(-3, 2))
torch.flip(input=my_tensor, dims=(-3, -1))
# tensor([[[8, 0, 5], [1, 6, 3]],
#         [[4, 7, 2], [2, 3, 8]]])

torch.flip(input=my_tensor, dims=(1, 2))
torch.flip(input=my_tensor, dims=(1, -1))
torch.flip(input=my_tensor, dims=(2, 1))
torch.flip(input=my_tensor, dims=(2, -2))
torch.flip(input=my_tensor, dims=(-1, 1))
torch.flip(input=my_tensor, dims=(-1, -2))
torch.flip(input=my_tensor, dims=(-2, 2))
torch.flip(input=my_tensor, dims=(-2, -1))
# tensor([[[2, 3, 8], [4, 7, 2]],
#         [[1, 6, 3], [8, 0, 5]]])

torch.flip(input=my_tensor, dims=(0, 1, 2))
etc.
# tensor([[[1, 6, 3], [8, 0, 5]],
#         [[2, 3, 8], [4, 7, 2]]])

my_tensor = torch.tensor([[[2., 7., 4.], [8., 3., 2.]], # 3D tensor
                          [[5., 0., 8.], [3., 6., 1.]]])
torch.flip(input=my_tensor, dims=(0,))
# tensor([[[5., 0., 8.], [3., 6., 1.]],
#         [[2., 7., 4.], [8., 3., 2.]]])

my_tensor = torch.tensor([[[2.+0.j, 7.+0.j, 4.+0.j], # 3D tensor
                           [8.+0.j, 3.+0.j, 2.+0.j]],
                          [[5.+0.j, 0.+0.j, 8.+0.j],
                           [3.+0.j, 6.+0.j, 1.+0.j]]])
torch.flip(input=my_tensor, dims=(0,))
# tensor([[[5.+0.j, 0.+0.j, 8.+0.j],
#          [3.+0.j, 6.+0.j, 1.+0.j]],
#         [[2.+0.j, 7.+0.j, 4.+0.j],
#          [8.+0.j, 3.+0.j, 2.+0.j]]])
                         # 3D tensor
my_tensor = torch.tensor([[[True, False, True], [True, False, True]],
                          [[False, True, False], [False, True, False]]])
torch.flip(input=my_tensor, dims=(0,))
# tensor([[[False, True, False], [False, True, False]],
#         [[True, False, True], [True, False, True]]])
Enter fullscreen mode Exit fullscreen mode

Postmark Image

Speedy emails, satisfied customers

Are delayed transactional emails costing you user satisfaction? Postmark delivers your emails almost instantly, keeping your customers happy and connected.

Sign up

Top comments (0)

Billboard image

The Next Generation Developer Platform

Coherence is the first Platform-as-a-Service you can control. Unlike "black-box" platforms that are opinionated about the infra you can deploy, Coherence is powered by CNC, the open-source IaC framework, which offers limitless customization.

Learn more

👋 Kindness is contagious

Please leave a ❤️ or a friendly comment on this post if you found it helpful!

Okay