Question about the example dataset

[HelloWorldLTY]

Hi, thanks for your interesting work and releasing UNI v2. I have a quick question about the exampel dataset (in assest/data/). I think each data here contains one patch from the original TCGA slide image, is it correct?

Moreover, I have tried using the UNI model to directly encode whole-slide level information, and the result seems good. Do you have any ideas about how to generate slide embeddings without splitting it into small patches? Thanks.