Ceil and Floor Methods in PyTorch

A ceiling value of a number is the smallest integer greater than or equal to the number. To find the ceiling of the elements of a torch tensor, we use the torch.ceil() function. This function takes a torch tensor as input parameter and returns a torch tensor with the ceil values of each element of the input tensor. This function supports only real-valued inputs. It supports torch tensors of any dimension.

A floor value of a number is the largest integer less than or equal to the number. To find the floor of the elements of a torch tensor, we use the torch.floor() function. This function takes a torch tensor as input parameter and returns a torch tensor with the floor values of each element of the input tensor. This function supports only real-valued input and it can support torch tensors of any dimension.

Syntax

torch.ceil(input)
torch.floor(input)

Parameters

• input - It's the input tensor.

Output

It returns a tensor of ceil or floor of elements of input tensor.

Steps

You could use the following steps to find the tensor with ceil or floor of elements of an input tensor

• Import the required library. In all the following examples, the required Python library is torch. Make sure you have already installed it.

import torch

• Create a torch tensor and print the tensor.

T = torch.tensor([1, 2.3, .2, 0, 4.01, 4.5, 4.99])
print("Tensor:
", T)

• Compute torch.ceil(input) or torch.floor(input) to find the tensor with ceil or floor of the elements of input and print the computed value

print("Tensor with ceil values:
",torch.ceil(T))
print("Tensor with floor values:
",torch.floor(T))

Example 1

In the following Python code, we compute the ceiling and floor values of the elements of a torch tensor. We have taken all the elements of the tensor as positive numbers.

# Import the required library
import torch

# define a torch tensor
T = torch.tensor([1, 2.3, .2, 0, 4.01, 4.5, 4.99])
print("Tensor:
", T)
print("Tensor with ceil values:
",torch.ceil(T))
print("Tensor with floor values:
",torch.floor(T))

Output

Tensor:
   tensor([1.0000, 2.3000, 0.2000, 0.0000, 4.0100, 4.5000, 4.9900])
Tensor with ceil values:
   tensor([1., 3., 1., 0., 5., 5., 5.])
Tensor with floor values:
   tensor([1., 2., 0., 0., 4., 4., 4.])

Example 2

The following Python program shows how to find the ceiling and floor of the elements of a tensor. Some elements are negative. Notice how the ceiling and floor values of negative numbers are computed.

# Import the required library
import torch
T = torch.tensor([-1, -2.3, .2, 0, -4.01, 4.5, -4.99])
print("Tensor:
", T)
print("Tensor with ceil values:
",torch.ceil(T))
print("Tensor with floor values:
",torch.floor(T))

Output

Tensor:
   tensor([-1.0000, -2.3000, 0.2000, 0.0000, -4.0100, 4.5000, -4.9900])
Tensor with ceil values:
   tensor([-1., -2., 1., 0., -4., 5., -4.])
Tensor with floor values:
   tensor([-1., -3., 0., 0., -5., 4., -5.])

Example 3

In the following Python code, the input is a two-dimensional tensor. It computes the ceiling and floor of each element of the tensor

# Import the required library
import torch
T = torch.randn(4,3)
print("Tensor:
", T)
print("Tensor with ceil values:
",torch.ceil(T))
print("Tensor with floor values:
",torch.floor(T))

Output

Tensor:
   tensor([[-0.4304, -0.5405, -0.7153],
      [ 0.8230, -0.0368, -0.0357],
      [-1.3842, 0.2168, -0.0332],
      [ 0.3007, 0.2878, 0.1758]])
Tensor with ceil values:
   tensor([[-0., -0., -0.],
      [ 1., -0., -0.],
      [-1., 1., -0.],
      [ 1., 1., 1.]])
Tensor with floor values:
   tensor([[-1., -1., -1.],
      [ 0., -1., -1.],
      [-2., 0., -1.],
      [ 0., 0., 0.]])