search

ocean-data-factory-sweden / kso

Notebooks to upload/download marine footage, connect to a citizen science project, train machine learning models and publish marine biological observations.
GNU General Public License v3.0
4 stars 12 forks source link

Requests dependencies issue in Colab again #232

Closed victor-wildlife closed 10 months ago

victor-wildlife commented 11 months ago

I am afraid the current dev version is still having the same "requests" depency issue we had a while back

victor-wildlife commented 11 months ago

Image

jannesgg commented 10 months ago

@victor-wildlife After looking into options to fix this, including trying to pin a version of google-colab to install, it seems that restarting the kernel is the only viable solution. I also looked into upgrading the version of panoptes, but even the latest version requires requests to be below 2.29.

My suggestion would be to tell users to restart the kernel if they encounter this error. Might be best to say "Restart the kernel and then run the "Import Python packages" cell again".

We might have to modify that cell a bit though so that it works after the restart, here is a proposal:

https://codefile.io/f/zB7OnjdjQU

What do you think?

victor-wildlife commented 10 months ago

Thanks @jannesgg I have coded a potential solution of the first cell https://codefile.io/f/t23FT730Vr but after I ran it successfully, I got the following torch error in the second cell. Any thoughts?

Image

victor-wildlife commented 10 months ago

Also, it seems like the panoptes package might solve our request conflict issues https://github.com/zooniverse/panoptes-python-client/pull/301

jannesgg commented 10 months ago

Hey @victor-wildlife, thanks for taking a look at this. Would be great if they pull through the version update :)

As for the error you are getting, it should not be going into the part "Installing latest version from PyPI". I found that by using the following for cell 2: https://codefile.io/f/fx4s1RgezT, it did not return any of these errors. Have a look and see what you think.