Improving the clothing image classifier with data augmentation

Armed with several common augmentation methods, we will now apply them to train our image classifier on a small dataset in the following steps:

1. We start by constructing the transform function by combining all the data augmentation techniques we just discussed:

   >>> torch.manual_seed(42)
   >>> transform_train = transforms.Compose([
                          transforms.RandomHorizontalFlip(),
                          transforms.RandomRotation(10),
                          transforms.RandomResizedCrop(size=(28, 28),
                                                       scale=(0.9, 1)),
                          transforms.ToTensor(),
       ])
   

Here, we employ horizontal flip, rotation of up to 10 degrees, and cropping, with dimensions ranging from 90% to 100% of the original size.

2. We reload the training dataset with this transform function and only use 500 samples for training...