search

AdalbertoCq / Histomorphological-Phenotype-Learning

Corresponding code of 'Quiros A.C.+, Coudray N.+, Yeaton A., Yang X., Chiriboga L., Karimkhan A., Narula N., Pass H., Moreira A.L., Le Quesne J.*, Tsirigos A.*, and Yuan K.* Mapping the landscape of histomorphological cancer phenotypes using self-supervised learning on unlabeled, unannotated pathology slides. 2024'
37 stars 8 forks source link

Simpler trained model for inference? #3

Closed stygian2a closed 6 months ago

stygian2a commented 2 years ago

Hi! Thank you for your wonderful work!

I'm trying to load your trained networks, but I'm having some difficulties. I managed to load the graph correctly, but I can't seem to find which tensors correspond to input/outputs... There are some gradients and Adam tensors and operations everywhere.

Would you think it would be possible for you to propose simpler weights/models if we are only looking to use it for inference? Or, alternatively, a small script/notebook that shows how to load and use the model (Not inside your particular pipeline, but on its own.)

Thank you!

AdalbertoCq commented 2 years ago

Hey @stygian2a,

Thanks for the feedback :)

If you just want to use the self-supervised pre-trained model for inference, it should be relatively easy to run step 3 - Tile vector representations of the flow to work with external cohorts.

That should give you vector representations for each tile image of your own WSIs.

Thanks, Adal