Deep Learning with PyTorch

Creating PyTorch datasets

We create a transformation object containing all the basic transformations required and use the ImageFolder to load the images from the data directory that we created in Chapter 5, Deep Learning for Computer Vision. In the following code, we create the datasets:

data_transform = transforms.Compose([        transforms.Resize((299,299)),        transforms.ToTensor(),        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])    ])# For Dogs & Cats datasettrain_dset = ImageFolder('../../chapter5/dogsandcats/train/',transform=data_transform)val_dset = ImageFolder('../../chapter5/dogsandcats/valid/',transform=data_transform)classes=2

By now, most of the preceding code will be self-explanatory.