jamesdolezal / slideflow

Deep learning library for digital pathology, with both Tensorflow and PyTorch support.
https://slideflow.dev
GNU General Public License v3.0
234 stars 39 forks source link

Generate features from SimCLR model #303

Closed matte-esse closed 11 months ago

matte-esse commented 11 months ago

I had to comment out as_pretrained in :meth:simclr.load

https://github.com/jamesdolezal/slideflow/blob/ea615e35ee07371da8afa12ff37b5b7166c1085f/slideflow/model/extractors/simclr.py#L52

and center_crop in :meth:simclr.data.get_preprocess_fn

https://github.com/jamesdolezal/slideflow/blob/ea615e35ee07371da8afa12ff37b5b7166c1085f/slideflow/model/extractors/simclr.py#L64

In order to extract features from a SimCLR model.

Should I remove those arguments and push changes? (If so I would need to create a new dev branch since I (rightfully) don't have permission to push changes to master.

jamesdolezal commented 11 months ago

Did you update the simclr submodule?

git submodule update --recursive

matte-esse commented 11 months ago

No, sorry. That was it