Data augmentation advice (horizontal flipping)

vision
1

transform = transforms.Compose([transforms.Resize((224, 244)),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize(mean = [0.5, 0.5, 0.5],
std = [0.5, 0.5, 0.5])])

this is what I did for my dataset.
before applying randomhorizontalflip, there was 2299 number of mini_batches. And after applying horizontal flipping there are still 2299 number of mini_batches. I printed out the len(data_loader).
why is this not working?

2

It is working, since this transformation will be randomly applied to each sample. The length of your Dataset won’t change, but in the default setup, each sample will be flipped with p=0.5.

3

thanks for the reply!! have a nice day :slight_smile: