Augumentation and agumentation data loader

Hi,

I have been checking how augmentation works in V3.
I like the whole concept how is it implemented :slight_smile:
But don’t understand basic how many arguments images is feed during training or validation in one cycle?

Thank

Michal