DEV Community

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

Posted on • Edited on

vstack and dstack in PyTorch

Buy Me a Coffee

*Memos:

vstack() can get the 1D or more D vertically(row-wisely) stacked tensor of zero or more elements from the one or more 0D or more D tensors of zero or more elements as shown below:

*Memos:

  • vstack() can be used with torch but not with a tensor.
  • The 1st argument with torch is tensors(Required-Type:tuple or list of tensor of int, float, complex or bool). *Basically, the size of tensors must be the same.
  • There is out argument with torch(Optional-Default:None-Type:tensor): *Memos:
    • out= must be used.
    • My post explains out argument.
  • row_stack() is the alias of vstack().
import torch

tensor1 = torch.tensor(2)
tensor2 = torch.tensor(7)
tensor3 = torch.tensor(4)

torch.vstack(tensors=(tensor1, tensor2, tensor3))
# tensor([[2], [7], [4]])

tensor1 = torch.tensor([2, 7, 4])
tensor2 = torch.tensor([8, 3, 2])
tensor3 = torch.tensor([5, 0, 8])

torch.vstack(tensors=(tensor1, tensor2, tensor3))
# tensor([[2, 7, 4], [8, 3, 2], [5, 0, 8]])

tensor1 = torch.tensor([[2, 7, 4], [8, 3, 2]])
tensor2 = torch.tensor([[5, 0, 8], [3, 6, 1]])
tensor3 = torch.tensor([[9, 4, 7], [1, 0, 5]])

torch.vstack(tensors=(tensor1, tensor2, tensor3))
# tensor([[2, 7, 4],
#         [8, 3, 2],
#         [5, 0, 8],
#         [3, 6, 1],
#         [9, 4, 7],
#         [1, 0, 5]])

tensor1 = torch.tensor([[2., 7., 4.], [8., 3., 2.]])
tensor2 = torch.tensor([[5., 0., 8.], [3., 6., 1.]])
tensor3 = torch.tensor([[9., 4., 7.], [1., 0., 5.]])

torch.vstack(tensors=(tensor1, tensor2, tensor3))
# tensor([[2., 7., 4.],
#         [8., 3., 2.],
#         [5., 0., 8.],
#         [3., 6., 1.],
#         [9., 4., 7.],
#         [1., 0., 5.]])

tensor1 = torch.tensor([[2.+0.j, 7.+0.j, 4.+0.j],
                        [8.+0.j, 3.+0.j, 2.+0.j]])
tensor2 = torch.tensor([[5.+0.j, 0.+0.j, 8.+0.j],
                        [3.+0.j, 6.+0.j, 1.+0.j]])
tensor3 = torch.tensor([[9.+0.j, 4.+0.j, 7.+0.j],
                        [1.+0.j, 0.+0.j, 5.+0.j]])
torch.vstack(tensors=(tensor1, tensor2, tensor3))
# tensor([[2.+0.j, 7.+0.j, 4.+0.j],
#         [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],
#         [9.+0.j, 4.+0.j, 7.+0.j],
#         [1.+0.j, 0.+0.j, 5.+0.j]])

tensor1 = torch.tensor([[True, False, True], [False, True, False]])
tensor2 = torch.tensor([[False, True, False], [True, False, True]])
tensor3 = torch.tensor([[True, False, True], [False, True, False]])

torch.vstack(tensors=(tensor1, tensor2, tensor3))
# tensor([[True, False, True],
#         [False, True, False],
#         [False, True, False],
#         [True, False, True],
#         [True, False, True],
#         [False, True, False]])

tensor1 = torch.tensor([[]])
tensor2 = torch.tensor([])
tensor3 = torch.tensor([[]])

torch.vstack(tensors=(tensor1, tensor2, tensor3))
# tensor([], size=(3, 0))
Enter fullscreen mode Exit fullscreen mode

dstack() can get the 3D or more D depth-wisely stacked tensor of zero or more elements from the one or more 0D or more D tensors of zero or more elements as shown below:

*Memos:

  • dstack() can be used with torch but not with a tensor.
  • The 1st argument with torch is tensors(Required-Type:tuple or list of tensor of int, float, complex or bool). *Basically, the size of tensors must be the same.
  • There is out argument with torch(Optional-Default:None-Type:tensor): *Memos:
    • out= must be used.
    • My post explains out argument.
import torch

tensor1 = torch.tensor(2)
tensor2 = torch.tensor(7)
tensor3 = torch.tensor(4)

torch.dstack(tensors=(tensor1, tensor2, tensor3))
# tensor([[[2, 7, 4]]])

tensor1 = torch.tensor([2, 7, 4])
tensor2 = torch.tensor([8, 3, 2])
tensor3 = torch.tensor([5, 0, 8])

torch.dstack(tensors=(tensor1, tensor2, tensor3))
# tensor([[[2, 8, 5], [7, 3, 0], [4, 2, 8]]])

tensor1 = torch.tensor([[2, 7, 4], [8, 3, 2]])
tensor2 = torch.tensor([[5, 0, 8], [3, 6, 1]])
tensor3 = torch.tensor([[9, 4, 7], [1, 0, 5]])

torch.dstack(tensors=(tensor1, tensor2, tensor3))
# tensor([[[2, 5, 9], [7, 0, 4], [4, 8, 7]],
#         [[8, 3, 1], [3, 6, 0], [2, 1, 5]]])

tensor1 = torch.tensor([[2., 7., 4.], [8., 3., 2.]])
tensor2 = torch.tensor([[5., 0., 8.], [3., 6., 1.]])
tensor3 = torch.tensor([[9., 4., 7.], [1., 0., 5.]])

torch.dstack(tensors=(tensor1, tensor2, tensor3))
# tensor([[[2., 5., 9.], [7., 0., 4.], [4., 8., 7.]],
#         [[8., 3., 1.], [3., 6., 0.], [2., 1., 5.]]])

tensor1 = torch.tensor([[2.+0.j, 7.+0.j, 4.+0.j],
                        [8.+0.j, 3.+0.j, 2.+0.j]])
tensor2 = torch.tensor([[5.+0.j, 0.+0.j, 8.+0.j],
                        [3.+0.j, 6.+0.j, 1.+0.j]])
tensor3 = torch.tensor([[9.+0.j, 4.+0.j, 7.+0.j],
                        [1.+0.j, 0.+0.j, 5.+0.j]])
torch.dstack(tensors=(tensor1, tensor2, tensor3))
# tensor([[[2.+0.j, 5.+0.j, 9.+0.j],
#          [7.+0.j, 0.+0.j, 4.+0.j],
#          [4.+0.j, 8.+0.j, 7.+0.j]],
#         [[8.+0.j, 3.+0.j, 1.+0.j],
#          [3.+0.j, 6.+0.j, 0.+0.j],
#          [2.+0.j, 1.+0.j, 5.+0.j]]])

tensor1 = torch.tensor([[True, False, True], [False, True, False]])
tensor2 = torch.tensor([[False, True, False], [True, False, True]])
tensor3 = torch.tensor([[True, False, True], [False, True, False]])

torch.dstack(tensors=(tensor1, tensor2, tensor3))
# tensor([[[True, False, True],
#          [False, True, False],
#          [True, False, True]],
#         [[False, True, False],
#          [True, False, True],
#          [False, True, False]]])

tensor1 = torch.tensor([[]])
tensor2 = torch.tensor([])
tensor3 = torch.tensor([[]])

torch.dstack(tensors=(tensor1, tensor2, tensor3))
# tensor([], size=(1, 0, 3))
Enter fullscreen mode Exit fullscreen mode

Sentry image

See why 4M developers consider Sentry, “not bad.”

Fixing code doesn’t have to be the worst part of your day. Learn how Sentry can help.

Learn more

Top comments (0)

Billboard image

Create up to 10 Postgres Databases on Neon's free plan.

If you're starting a new project, Neon has got your databases covered. No credit cards. No trials. No getting in your way.

Try Neon for Free →

👋 Kindness is contagious

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

Okay