How to Make a grid of Images in PyTorch?

In this article, we are going to see How to Make a grid of Images in PyTorch. we can make a grid of images using the make_grid() function of torchvision.utils package.

make_grid() function:

The make_grid() function accept 4D tensor with [B, C ,H ,W] shape. where B represents the batch size, C represents the number of channels, and  H, W represents the height and width respectively. The height and weight should be the same for all images. This function returns the tensor that contains a grid of input images. we can also set the number of images displayed in each row by using nrow parameter. The below syntax is used to make a grid of images in PyTorch.

Syntax: torchvision.utils.make_grid(tensor)

Parameter:

• tensor (Tensor or list) tensor of shape (B x C x H x W) or a list of images all of the same size.
• nrow (int, optional) – Number of images displayed in each row of the grid. Default: 8.
• padding (int, optional) – amount of padding. Default: 2.

Returns: This function returns the tensor that contains a grid of input images.

Example 1:

The following example is to understand how to make a grid of images in PyTorch.

Python3

# import required library
import torch
import torchvision
from torchvision.io import read_image
from torchvision.utils import make_grid

# read images from computer
a = read_image('a.jpg')
b = read_image('b.jpg')
c = read_image('c.jpg')
d = read_image('d.jpg')

# make grid from the input images
# this grid contain 4 columns and 1 row
Grid = make_grid([a, b, c, d])

# display result
img = torchvision.transforms.ToPILImage()(Grid)
img.show()

Output:

Example 2:

in the following example, we make a grid of images and set the number of images displayed in each row by using nrow Parameter.

Python3

# import required library
import torch
import torchvision
from torchvision.io import read_image
from torchvision.utils import make_grid

# read images from computer
a = read_image('a.jpg')
b = read_image('b.jpg')
c = read_image('c.jpg')
d = read_image('d.jpg')
e = read_image('e.jpg')
f = read_image('f.jpg')

# make grid from the input images
# this grid contain 2 rows and 3 columns
Grid = make_grid([a, b, c, d, e, f], nrow=3)

# display result
img = torchvision.transforms.ToPILImage()(Grid)
img.show()

Output:

Example 3:

In the following example, we make a grid of images and set the padding between the images.

Python3

# import required library
import torch
import torchvision
from torchvision.io import read_image
from torchvision.utils import make_grid

# read images from computer
a = read_image('a.png')
b = read_image('b.png')
c = read_image('c.png')
d = read_image('d.png')
e = read_image('e.png')
f = read_image('f.png')

# make grid from the input images
# set nrow=3, and padding=25
Grid = make_grid([a, b, c, d, e, f], nrow=3, padding=25)

# display result
img = torchvision.transforms.ToPILImage()(Grid)
img.show()

Output: