tarrinw
commented
3 months ago Something was wrong with the nvidia libraries. Solution: upgrade / force reinstall of these libraries:
pip install nvidia-cufft-cu12 -U
pip install nvidia-cusparse-cu11 -U
pip install nvidia-cusparse-cu12 -U
Using kraken 4.3.13 segmentation now fails due to dependencies on two nvidia libraries: libcurand and libcufft
Previously all operations could run without gpu and gpu libraries.
Traceback (most recent call last): File "/usr/local/lib64/python3.9/site-packages/torch/init.py", line 168, in _load_global_deps ctypes.CDLL(lib_path, mode=ctypes.RTLD_GLOBAL) File "/usr/lib64/python3.9/ctypes/init.py", line 374, in init self._handle = _dlopen(self._name, mode) OSError: libcurand.so.10: cannot open shared object file: No such file or directory
During handling of the above exception, another exception occurred:
Traceback (most recent call last): File "/usr/local/bin/kraken", line 5, in
from kraken.kraken import cli
File "/usr/local/lib/python3.9/site-packages/kraken/kraken.py", line 35, in
from kraken.lib.progress import KrakenProgressBar, KrakenDownloadProgressBar
File "/usr/local/lib/python3.9/site-packages/kraken/lib/progress.py", line 22, in
import pytorch_lightning as pl
File "/usr/local/lib/python3.9/site-packages/pytorch_lightning/init.py", line 25, in
from lightning_fabric.utilities.seed import seed_everything # noqa: E402
File "/usr/local/lib/python3.9/site-packages/lightning_fabric/init.py", line 29, in
from lightning_fabric.fabric import Fabric # noqa: E402
File "/usr/local/lib/python3.9/site-packages/lightning_fabric/fabric.py", line 21, in
import torch
File "/usr/local/lib64/python3.9/site-packages/torch/init.py", line 228, in
_load_global_deps()
File "/usr/local/lib64/python3.9/site-packages/torch/init.py", line 190, in _load_global_deps
ctypes.CDLL(lib_path, mode=ctypes.RTLD_GLOBAL)
File "/usr/lib64/python3.9/ctypes/init.py", line 374, in init
self._handle = _dlopen(self._name, mode)
OSError: libcufft.so.10: cannot open shared object file: No such file or directory