Open ManuelNickel opened 1 month ago
Thanks @ManuelNickel !
For me cloning gandalf wasnt enough, I cloned gandalf and ended with a error that said
ERROR: Could not find a version that satisfies the requirement torch==1.8.2 (from gandlf) (from versions: 1.7.1, 1.8.0, 1.8.1, 1.9.0, 1.9.1, 1.10.0, 1.10.1, 1.10.2, 1.11.0, 1.12.0, 1.12.1, 1.13.0, 1.13.1, 2.0.0, 2.0.1, 2.1.0, 2.1.1, 2.1.2, 2.2.0, 2.2.1, 2.2.2, 2.3.0, 2.3.1) ERROR: No matching distribution found for torch==1.8.2
I found the file with this requirement, located in
Challenge\Task_1\GaNDLF\tutorials\classification_pathmnist_notebook\classification_tutorial.ipynb
and changed it to 1.8.1
Was finally able to install GANDLF using pip. Putting this here incase any one else goes through the same error
Also, As above mentioned, Clone Algorithms (comment them out from setup.py), and renaming GANDALF_modules to GANDALF and then installing with pip, worked for me @ManuelNickel I guess thats what you suggested as well
Dear FeTS-AI team,
following the instructions on how to install the infrastructure for Task 1, I encountered the following error for step 8, the pip install step:
Looking at Algorithms/.gitmodules, I noticed that the GANDLF submodule has been renamed to GANDLF_module a while back.
I got around this by manually cloning openfl, GANDFL and fets from the repos as listed in
Challenge/Task_1/setup.py
within theinstall_requires
block, checking out the corresponding commits andpip install
ing these into the env. Then, I commented out theinstall_requires
block and ranpip install .
onChallenge/Task_1
again.My conda env was then still missing the correct GLIB version, erroring out with
libstdc++.so.6: version GLIBCXX_3.4.30' not found
.conda install -c conda-forge libgcc=5.2.0
fixed this for me.Lastly, when installing
torchvision
, one needs to pin it to an older version, as the newer ones do not provide the requiredtorch.ao
module:So, easy fix. Just leave this here in case someone else stumbles upon this. 🙂
Best, Manu