I'm working on an image classification using PyTorch. My dataset is in directory format. I've set up my data pipeline and model. Still, I'm encountering an issue with my train-validation split where some classes have zero samples in either the training or validation datasets. Here's the relevant part of my…