You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The reason will be displayed to describe this comment to others. Learn more.
The default Git runner for ubuntu-latest does not come with CuDA. Unfortunately, the default pip install torch on Ubuntu installs the CuDA version, which gives the error OSError: libcufft.so.11: cannot open shared object file: No such file or directory when running pytest in CI. This can probably be fixed by installing the CPU only version via pip install torch --index-url https://download.pytorch.org/whl/cpu (or even a version with different CuDA via pip install torch --index-url https://download.pytorch.org/whl/cu118), but I don't know how to give pip install options to poetry (the poetry documentation is not clear on this). Hence, I'm switching the tests to run on MacOS for now.
Note that I ran all the tests via pytest on an Ubuntu machine (Galois Razor), and they all pass. However, this was after I installed the proper version of CuDA on this Ubuntu machine by passing options to pip install. So, I'm confident that the code is fine and will run properly on Linux (or Ubuntu at least), but I have to make the Github runner Mac to make CI pass due to peculiarities of poetry and Git runners.
f9f764dThere was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The default Git runner for
ubuntu-latestdoes not come with CuDA. Unfortunately, the defaultpip install torchon Ubuntu installs the CuDA version, which gives the errorOSError: libcufft.so.11: cannot open shared object file: No such file or directorywhen runningpytestin CI. This can probably be fixed by installing the CPU only version viapip install torch --index-url https://download.pytorch.org/whl/cpu(or even a version with different CuDA viapip install torch --index-url https://download.pytorch.org/whl/cu118), but I don't know how to give pip install options to poetry (the poetry documentation is not clear on this). Hence, I'm switching the tests to run on MacOS for now.Note that I ran all the tests via
pyteston an Ubuntu machine (Galois Razor), and they all pass. However, this was after I installed the proper version of CuDA on this Ubuntu machine by passing options topip install. So, I'm confident that the code is fine and will run properly on Linux (or Ubuntu at least), but I have to make the Github runner Mac to make CI pass due to peculiarities of poetry and Git runners.