-
Notifications
You must be signed in to change notification settings - Fork 15
Open
Labels
bugSomething isn't workingSomething isn't working
Description
Using the N2V3d notebook, I can train a model using my GPU and check that it works; it does! I would like to be able to use it in ImageJ, and the .ipynb says in section 3.2 that this export should happen automatically. It does not. When I look in the model directory I can see the two weights.h5 files, config.json, Quality control, logs directories and a training report pdf, but no .zip file.
To Reproduce
Run DlC4Mic docker container with N2V 3D model, then train on suitable dataset.
Expected behavior
Creation of bioimage model zoo model
I'm leaving out a lot of information like CUDA version etc. as it seems like the ipynb generally works, so I don't think that's the problem.
Metadata
Metadata
Assignees
Labels
bugSomething isn't workingSomething isn't working