Hello
I am currently on the end phase of my Capstone Project from my DAS and I used to use 2 GPUs amd Parallel computing since I ran into a memory error with 1 GPU before. I think there was an update recently and now, amongst some other things that just changed and I needed to fix, I can’t use 2 GPUs anymore. Is there anyway to get them back, for the next few weeks?
Best
Patrick
Hi Patrick,
the two GPUs that were available previously were very old/outdated P100s. Those are not available anymore on the new infrastructure. You should now have access to a 20GB partition of an A100 - have you tried running your code there?
What else did you need to fix?
Thank you for the answer. I will adapt the code, I think that should work aswell.
The error looked like this:
raceback (most recent call last):
File “/home/renku/work/GLoG-CSUnet/train_CSUnet_HVSMR.py”, line 15, in
from utils.utils import DiceLoss, cleanup_old_models, add_results_to_experimets_file
File “/home/renku/work/GLoG-CSUnet/utils/utils.py”, line 15, in
import cv2
File “/home/renku/work/.venv/lib/python3.13/site-packages/cv2/init.py”, line 181, in
bootstrap()
~~~~~~~~~^^
File “/home/renku/work/.venv/lib/python3.13/site-packages/cv2/init.py”, line 153, in bootstrap
native_module = importlib.import_module("cv2")
File “/layers/paketo-buildpacks_cpython/cpython/lib/python3.13/importlib/init.py”, line 88, in import_module
return _bootstrap._gcd_import(name[level:], package, level)
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
ImportError: libGL.so.1: cannot open shared object file: No such file or directory
I got rid of it by removing the import. Further the path (seen in the error aswell) used to be /home/jovyan/… and not /home/renku/…
We updated the images for the global environments because they were getting outdated. You can keep using the old image, but you need to set up a custom launcher. Have a look at this one for an example on how to set it up.