CellFinder training

[danifranco]

Hello!

Great work what you have done here!

I'm curious how you trained CellFinder model. In the paper you mentioned to have used several open-source datasets of different image modalities together with a private dataset Phase400. Did you train a single CellFinder/AnchorDETR model with all data at once or one model for each image type/modality?

Are you going to release the training scheme of CellFinder? (related to #18)

Thank you so much!