DEV Community

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

Posted on

squeeze in PyTorch

Buy Me a Coffee

*My post explains unsqueeze().

squeeze() can get the 0D or more D tensor of the zero or more elements whose zero or more dimensions are removed if the size is 1 from the 0D or more D tensor of zero or more elements as shown below:

*Memos:

  • squeeze() 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 dim(Optional-Type:int, tuple of int or list of int): *Memos:
    • Each number must be unique.
    • It can remove the specific zero or more dimensions whose size is 1.
    • If the size is not 1, zero or more dimensions are not removed even if you set it.
import torch

my_tensor = torch.tensor([[[[0], [1]],
                           [[2], [3]],
                           [[4], [5]]]])
torch.squeeze(input=my_tensor)
my_tensor.squeeze()
torch.squeeze(input=my_tensor, dim=(0, 3))
my_tensor.squeeze(dim=(0, 3))
my_tensor.squeeze(0, 3)
torch.squeeze(input=my_tensor, dim=(0, 1, 3))
my_tensor.squeeze(dim=(0, 1, 3))
my_tensor.squeeze(0, 1, 3)
etc.
torch.squeeze(input=my_tensor, dim=(0, 1, 2, 3))
my_tensor.squeeze(dim=(0, 1, 2, 3))
my_tensor.squeeze(0, 1, 2, 3)
etc.
# tensor([[0, 1],
#         [2, 3],
#         [4, 5]])

torch.squeeze(input=my_tensor, dim=0)
torch.squeeze(input=my_tensor, dim=-4)
torch.squeeze(input=my_tensor, dim=(0,))
torch.squeeze(input=my_tensor, dim=(-4,))
torch.squeeze(input=my_tensor, dim=(0, 1))
torch.squeeze(input=my_tensor, dim=(0, 2))
torch.squeeze(input=my_tensor, dim=(0, -2))
torch.squeeze(input=my_tensor, dim=(0, -3))
torch.squeeze(input=my_tensor, dim=(1, 0))
etc.
torch.squeeze(input=my_tensor, dim=(0, 1, 2))
etc.
# tensor([[[0], [1]],
#         [[2], [3]],
#         [[4], [5]]])

torch.squeeze(input=my_tensor, dim=1)
torch.squeeze(input=my_tensor, dim=2)
torch.squeeze(input=my_tensor, dim=-2)
torch.squeeze(input=my_tensor, dim=-3)
torch.squeeze(input=my_tensor, dim=())
torch.squeeze(input=my_tensor, dim=(1,))
torch.squeeze(input=my_tensor, dim=(2,))
torch.squeeze(input=my_tensor, dim=(-2,))
torch.squeeze(input=my_tensor, dim=(-3,))
torch.squeeze(input=my_tensor, dim=(1, 2))
etc.
# tensor([[[[0], [1]],
#          [[2], [3]],
#          [[4], [5]]]])

torch.squeeze(input=my_tensor, dim=3)
torch.squeeze(input=my_tensor, dim=-1)
torch.squeeze(input=my_tensor, dim=(3,))
torch.squeeze(input=my_tensor, dim=(-1,))
torch.squeeze(input=my_tensor, dim=(1, 3))
torch.squeeze(input=my_tensor, dim=(1, -1))
torch.squeeze(input=my_tensor, dim=(2, 3))
torch.squeeze(input=my_tensor, dim=(2, -1))
torch.squeeze(input=my_tensor, dim=(3, 1))
etc.
torch.squeeze(input=my_tensor, dim=(1, 2, 3))
etc.
# tensor([[[0, 1],
#          [2, 3],
#          [4, 5]]])

my_tensor = torch.tensor([[[[0.], [1.]],
                           [[2.], [3.]],
                           [[4.], [5.]]]])
torch.squeeze(input=my_tensor)
# tensor([[0., 1.],
#         [2., 3.],
#         [4., 5.]])

my_tensor = torch.tensor([[[[0.+0.j], [1.+0.j]],
                           [[2.+0.j], [3.+0.j]],
                           [[4.+0.j], [5.+0.j]]]])
torch.squeeze(input=my_tensor)
# tensor([[0.+0.j, 1.+0.j],
#         [2.+0.j, 3.+0.j],
#         [4.+0.j, 5.+0.j]])

my_tensor = torch.tensor([[[[True], [False]],
                           [[False], [True]],
                           [[True], [False]]]])
torch.squeeze(input=my_tensor)
# tensor([[True, False],
#         [False, True],
#         [True, False]])
Enter fullscreen mode Exit fullscreen mode

Sentry image

Hands-on debugging session: instrument, monitor, and fix

Join Lazar for a hands-on session where you’ll build it, break it, debug it, and fix it. You’ll set up Sentry, track errors, use Session Replay and Tracing, and leverage some good ol’ AI to find and fix issues fast.

RSVP here →

Top comments (0)

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

👋 Kindness is contagious

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

Okay