Torchvision.transforms

import torch
import torchvision.transforms.functional as TF
import random

def synchronized_transform(image, mask): 

    # Randomly choose the start point for the crop 
    i, j, h, w = TF.RandomCrop.get_params(image,

                                          output_size=(256, 256)) 

    image_cropped = TF.crop(image, i, j, h, w) 
    mask_cropped = TF.crop(mask, i, j, h, w) 

    # Apply any other synchronized transformations
here 

    if random.random() > 0.5: 

        image_cropped = TF.hflip(image_cropped) 

        mask_cropped = TF.hflip(mask_cropped) 
    return image_cropped, mask_cropped

Saliency map

import torch
from torchvision import models, transforms
from PIL import Image

# Load model and image
model = models.resnet50(pretrained=True).eval()
img = Image.open("path/to/image.jpg")

# Preprocess image
prep = transforms.Compose([
 transforms.Resize(256), 
 transforms.CenterCrop(224),
 transforms.ToTensor(), 
 transforms.Normalize([0.485, 0.456, 0.406], 
[0.229, 0.224, 0.225])
])

img_t = prep(img).unsqueeze(0).requires_grad_()

# Forward pass
output = model(img_t)
output[0, output.argmax()].backward()

# Saliency map
saliency = img_t.grad.data.abs().squeeze()