sample_sanity_check.py

import torchvision.transforms as transforms
import torchvision.datasets as datasets
import torchvision
from modules.cifar10 import data_loader
import matplotlib.pyplot as plt

# modules.utils.py
class DeNormalize(object):
    def __init__(self, mean, std):
        self.mean = mean
        self.std = std
    def __call__(self, tensor):
        """
        Args:
            tensor (Tensor): Tensor image of size (C, H, W) to be normalized.
        Returns:
            Tensor: Normalized image.
        """
        for t, m, s in zip(tensor, self.mean, self.std):
            t.mul_(s).add_(m)
            # The normalize code -> t.sub_(m).div_(s)
        return tensor
 
# modules.cifar10.py
def denormalize_transform():
    denormal = DeNormalize(
        mean=[0.4914, 0.4822, 0.4465], std=[0.2023, 0.1994, 0.2010])
    return denormal

# Sanity Check in a notebook cell
train_loader, val_loader = data_loader('./data', batch_size=3)
(image, target) = iter(train_loader).next()

# Denormalize RGB values since our data loader has a normalize
# Values above or below [0,1] are clipped in the RGB image displayed
transform = denormalize_transform()
transform(image)
grid_img = torchvision.utils.make_grid(image, nrow=3)
plt.imshow(grid_img.permute(1, 2, 0))