MIL with Stain Augmentation

[sanosano]

@jamesdolezal
Thank you for your excellent and useful project.

I have a question regarding the application of stain augmentation.

In my understanding, the code mentioned in the document at https://slideflow.dev/norm/#stain-augmentation is used during the training of a tile-based model.

Is it possible to apply Stain Augmentation when training a multiple-instance learning model, as in the tutorial at https://slideflow.dev/tutorial8/# ?
The tutorial completes the creation of the feature bag before calling the train_mil function, but I think it is necessary to generate a feature bag with Stain Augmentation applied at each training iteration.
Does Slideflow support such a pipeline?