-
Notifications
You must be signed in to change notification settings - Fork 32
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Adapt download_model script to fetch release assets #797
Comments
One annoying behavior I noticed is related to our dual use of final and best checkpoints. (See axondeepseg/nn-axondeepseg#8) By default, nnunet looks for the final checkpoints but many of our models are the best ones. To use them, we need to specify to the
Note that the function mentioned above is used during inference, not when downloading the models, so we could also add a function that verifies the checkpoint type in I would actually prefer to not have a |
Describe the problem
The download_model.py script currently takes models by downloading the whole repository and extracting the model folder. It seems we used to put the
.pt
checkpoint files directly in the repo, but we don't do this anymore. Instead, we put them as release assets, so this method will not work anymore.Proposed solution
I propose we simply re-write the script. If users want to download default ivadomed models, they can always get an older version of ADS and run the script, which should still work in the future.
I already have a working script we can use here: https://github.com/axondeepseg/nn-axondeepseg/blob/main/download_models.py
Related to #765
The text was updated successfully, but these errors were encountered: