For which purpose is use_batch_labels used for?

[cam1flag]

Hi, I see in the code (e.g. get_all_embeddings_metric in train-saturn.py) there is this parameter use_batch_labels. It is described as 'if we add batch labels as a categorical covariate' I do not get the difference if I assign it to True or False. Could you calrify?

Thanks :)

[Yanay1]

use_batch_labels pairs with the argument non_species_batch_col. If you want to add an additional column from each dataset as a categorical covariate during pretraining of the autoencoder, you can pass the column name as that argument.