Should we be using dropbox for this? Perhaps GIT LFS through github makes more sense. Also using the program tar makes this almost impossible to run on Windows, Python's tarfile library perhaps should be used to access the contents.

thanks, updated to use tarfile. i don't have any preference for the file hosting please feel free to change

I agree that using git lfs with no tarball/zip or similar would be much better.

Should we be using dropbox for this? Perhaps GIT LFS through github makes more sense. Also using the program tar makes this almost impossible to run on Windows, Python's tarfile library perhaps should be used to access the contents.

Hi @ericspod ,

Could you please help set the expected storage for this dataset? Or feel free to submit PR directly.
Both Wenqi and I don't have much experience with it.
Thanks.

Hi @ericspod and @tvercaut ,

As we are closing to v0.1 release deadline, if we don't have a quick solution here, can we merge this PR first? At least users can get the link in the tutorial.
Thanks.

The simplest would be to create a new MedNIST (or similar name) repo under Project-MONAI and enable git lfs on it (see https://help.github.com/en/github/managing-large-files). That said, it would be best to do this once the licensing question discussed in #237 has been resolved.

I would leave sorting LFS off until after the v0.1 release. It's not critical, just a best practice thing, and will make more sense the more examples with large scale data we add.

I would leave sorting LFS off until after the v0.1 release. It's not critical, just a best practice thing, and will make more sense the more examples with large scale data we add.

So do you agree to merge this PR first for v0.1 and add LFS support later?
Thanks.

@tvercaut mentioned a licensing issue, once that is resolved yes we merge.