Pad in PyTorch

#python #pytorch #pad #v2

Buy Me a Coffee☕

*Memos:

My post explains OxfordIIITPet().

Pad() can add padding to zero or more images as shown below:

*Memos:

The 1st argument for initialization is padding(Required-Type:int or tuple/list(int)): *Memos:
- It can do add padding.
- A tuple/list must be the 1D with 2 or 4 elements.
The 2nd argument for initialization is fill(Optional-Default:0-Type:int, float or tuple/list(int or float)): *Memos:
- It can change the background of images. *The background can be seen when adding padding for images.
- A tuple/list should be the 1D with 3 elements.
The 3rd argument for initialization is padding_mode(Optional-Default:'constant'-Type:str). *'constant', 'edge', 'reflect' or 'symmetric' can be set to it.
The 1st argument is img(Required-Type:PIL Image or tensor(int)): *Memos:
- It must be a 3D or more D tensor.
- Don't use img=.
v2 is recommended to use according to V1 or V2? Which one should I use?.

from torchvision.datasets import OxfordIIITPet
from torchvision.transforms.v2 import Pad

pad = Pad(padding=100)
pad = Pad(padding=100, fill=0, padding_mode='constant')

pad
# Pad(padding=100, fill=0, padding_mode=constant)

pad.padding
# 100

pad.fill
# 0

pad.padding_mode
# 'constant'

origin_data = OxfordIIITPet(
    root="data",
    transform=None
    # transform=Pad(padding=0)
)

p50_data = OxfordIIITPet(
    root="data",
    transform=Pad(padding=50)
)

p100_data = OxfordIIITPet(
    root="data",
    transform=Pad(padding=100)
)

p150_data = OxfordIIITPet(
    root="data",
    transform=Pad(padding=150)
)

m50_data = OxfordIIITPet(
    root="data",
    transform=Pad(padding=-50)
)

m100_data = OxfordIIITPet(
    root="data",
    transform=Pad(padding=-100)
)

m150_data = OxfordIIITPet(
    root="data",
    transform=Pad(padding=-150)
)

p100p50_data = OxfordIIITPet(
    root="data",
    transform=Pad(padding=[100, 50])
)

m100m50_data = OxfordIIITPet(
    root="data",
    transform=Pad(padding=[-100, -50])
)

p100m50_data = OxfordIIITPet(
    root="data",
    transform=Pad(padding=[100, -50])
)

p25p50p75p100_data = OxfordIIITPet(
    root="data",
    transform=Pad(padding=[25, 50, 75, 100])
)

m25m50m75m100_data = OxfordIIITPet(
    root="data",
    transform=Pad(padding=[-25, -50, -75, -100])
)

p25m50p75m100_data = OxfordIIITPet(
    root="data",
    transform=Pad(padding=[25, -50, 75, -100])
)

p100fillgray_data = OxfordIIITPet(
    root="data",
    transform=Pad(padding=100, fill=150)
)

p100fillpurple_data = OxfordIIITPet(
    root="data",
    transform=Pad(padding=100, fill=[160, 32, 240])
)

p100edge_data = OxfordIIITPet(
    root="data",
    transform=Pad(padding=100, padding_mode="edge")
)

p100reflect_data = OxfordIIITPet(
    root="data",
    transform=Pad(padding=100, padding_mode="reflect")
)

p100symmetric_data = OxfordIIITPet(
    root="data",
    transform=Pad(padding=100, padding_mode="symmetric")
)

import matplotlib.pyplot as plt

def show_images1(data, main_title=None):
    plt.figure(figsize=(10, 5))
    plt.suptitle(t=main_title, y=0.8, fontsize=14)
    for i, (im, _) in zip(range(1, 6), data):
        plt.subplot(1, 5, i)
        plt.imshow(X=im)
        plt.xticks(ticks=[])
        plt.yticks(ticks=[])
    plt.tight_layout()
    plt.show()

show_images1(data=origin_data, main_title='origin_data')
show_images1(data=p50_data, main_title='p50_data')
show_images1(data=p100_data, main_title='p100_data')
show_images1(data=p150_data, main_title='p150_data')
print()
show_images1(data=origin_data, main_title='origin_data')
show_images1(data=m50_data, main_title='m50_data')
show_images1(data=m100_data, main_title='m100_data')
show_images1(data=m150_data, main_title='m150_data')
print()
show_images1(data=origin_data, main_title='origin_data')
show_images1(data=p100p50_data, main_title='p100p50_data')
show_images1(data=m100m50_data, main_title='m100m50_data')
show_images1(data=p100m50_data, main_title='p100m50_data')
print()
show_images1(data=origin_data, main_title='origin_data')
show_images1(data=p25p50p75p100_data, main_title='p25p50p75p100_data')
show_images1(data=m25m50m75m100_data, main_title='m25m50m75m100_data')
show_images1(data=p25m50p75m100_data, main_title='p25m50p75m100_data')
print()
show_images1(data=p100fillgray_data, main_title='p100fillgray_data')
show_images1(data=p100fillpurple_data, main_title='p100fillpurple_data')
print()
show_images1(data=p100edge_data, main_title='p100edge_data')
show_images1(data=p100reflect_data, main_title='p100reflect_data')
show_images1(data=p100symmetric_data, main_title='p100symmetric_data')

# ↓ ↓ ↓ ↓ ↓ ↓ The code below is identical to the code above. ↓ ↓ ↓ ↓ ↓ ↓
def show_images2(data, main_title=None, p=0, f=0, pm='constant'):
    plt.figure(figsize=(10, 5))
    plt.suptitle(t=main_title, y=0.8, fontsize=14)
    for i, (im, _) in zip(range(1, 6), data):
        plt.subplot(1, 5, i)
        pad = Pad(padding=p, fill=f, padding_mode=pm) # Here
        plt.imshow(X=pad(im)) # Here
        plt.xticks(ticks=[])
        plt.yticks(ticks=[])
    plt.tight_layout()
    plt.show()

show_images2(data=origin_data, main_title='origin_data')
show_images2(data=origin_data, main_title='p50_data', p=50)
show_images2(data=origin_data, main_title='p100_data', p=100)
show_images2(data=origin_data, main_title='p150_data', p=150)
print()
show_images2(data=origin_data, main_title='origin_data')
show_images2(data=origin_data, main_title='m50_data', p=-50)
show_images2(data=origin_data, main_title='m100_data', p=-100)
show_images2(data=origin_data, main_title='m150_data', p=-150)
print()
show_images2(data=origin_data, main_title='origin_data')
show_images2(data=origin_data, main_title='p100p50_data', p=[100, 50])
show_images2(data=origin_data, main_title='m100m50_data', p=[-100, -50])
show_images2(data=origin_data, main_title='p100m50_data', p=[100, -50])
print()
show_images2(data=origin_data, main_title='origin_data')
show_images2(data=origin_data, main_title='p25p50p75p100_data',
             p=[25, 50, 75, 100])
show_images2(data=origin_data, main_title='m25m50m75m100_data',
             p=[-25, -50, -75, -100])
show_images2(data=origin_data, main_title='p25m50p75m100_data',
             p=[25, -50, 75, -100])
print()
show_images2(data=origin_data, main_title='p100fillgray_data', p=100,
             f=[150])
show_images2(data=origin_data, main_title='p100fillpurple_data', p=100,
             f=[160, 32, 240])
print()
show_images2(data=origin_data, main_title='p100edge_data', p=100, 
             pm='edge')
show_images2(data=origin_data, main_title='p100reflect_data', p=100,
             pm='reflect')
show_images2(data=origin_data, main_title='p100symmetric_data', p=100,
             pm='symmetric')