PyTorch transforms. See this for saving images: https://gist.github.com/DanielTakeshi/bbaf432347aafa2e9878e93fd6982fd7

gistfile1.txt

I am testing different PyTorch transforms on my data, and saving sample
images as they are loaded during training. See examples below.

transforms.RandomResizedCrop(224)

https://pytorch.org/docs/stable/torchvision/transforms.html#torchvision.transforms.RandomResizedCrop

transforms.Resize(256)
transforms.CenterCrop(224)

There's no randomness here.

transforms.Resize(256)
transforms.RandomResizedCrop(224)

By default, the random resized crop might have an initial crop as small as 8 percent of the original image, as you can see by some of the "zoomed in" stuff. And from the other one we did earlier (the Resize just made this 256x256 here before applying the RandomResizedCrop).

transforms.RandomResizedCrop(224, scale=(0.8, 1.0))

Now we don't do the initial scaling. This is probably better for the data we have, less "zoomed in" stuff.

175             transforms.RandomResizedCrop(224, scale=(0.8, 1.0))
176             transforms.RandomVerticalFlip()

So when they say vertical flip, that means a flip ABOUT THE HORIZONTAL AXIS. It is more intuitive their way.

TL;DR in my code use the HORIZONTAL TRANSFORM, so we flip about the vertical axis.