mahmoodlab / panther Goto Github PK

Hello,
there are no files "create_downstream_model", "prepare_emb" in mil_models folder, line 21. from mil_models import create_downstream_model, prepare_emb in the trainer.py file.

Hi Song,

I appreciate you open source the code for this incredible work.

I wondering the how the feature directory look like. Specifically, what's FEAT_DIR_1? Is that all patch features in a certain WSI?

Hi, Song. Congratulations on your team's impressive work.

It seems that the subscript index (1, 2, 3) and the prototype color do not match in Figure 2 (Overview of PANTHER workflow).

Hi Song,

I don't know how you collect the csv file in the survival splits.

For example, in the TCGA-BRCA study, TCGA-A1-A0SK with os_censorship 0 and os_survival_days 967 cannot be found in the website page of this case.

Can you share the information to collect those survival information? Since it's not described in the paper as well.

Thanks!

Hi，how do I know if the Whole Slide Images (WSIs) I am using are at 20× magnification?

Hello! Awesome work. I am curious, if in the patching/feature extraction steps I'm using overlapping tiles, let's say tile size 224 and step size 112, how would the visualization work? Is that even possible? In CLAM it seems overlapping tiles can still be used when visualizing the attention scores, I'm wondering if the same thing can be done with PANTHER. Thanks!

Hi, in the corresponding paper, you benchmark PANTHER against DeepSets which you included DeepSets as an unsupervised approach. Since I do not see any of the code for it, I am curious how you used DeepSets in an unsupervised fashion since my understanding is that the original paper uses it in a purely supervised fashion. Thanks!