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
'# 'Pass --no-use-pep517 to speed up rebuild by using the legacy build system
'# 'which doesn't use a one-time tmp directory for the build
pip3 install -e '.[dev]' --no-build-isolation
'#' Or:
'# 'python3 setup.py develop
'#'Code formatting and linting
pip3 install pre-commit
pre-commit install
pre-commit run --all-files
after changing python to 3.10, error fixed. But now i get Could not load library libnvrtc.so.12. Error: libnvrtc.so.12: cannot open shared object file: No such file or directory
Could not load library libnvrtc.so. Error: libnvrtc.so: cannot open shared object file: No such file or directory
Could not load library libnvrtc.so. Error: libnvrtc.so: cannot open shared object file: No such file or directory . what version uf cuda should I use? mlspace environments create --env name_test --python 3.10 --cuda *
kazakovaanastasia
changed the title
ModuleNotFoundError: No module named 'joblib'
Error: libnvrtc.so.12: cannot open shared object file: No such file or directory
Jan 12, 2025
after changing python to 3.10, error fixed. But now i get Could not load library libnvrtc.so.12. Error: libnvrtc.so.12: cannot open shared object file: No such file or directory
Could not load library libnvrtc.so. Error: libnvrtc.so: cannot open shared object file: No such file or directory
Could not load library libnvrtc.so. Error: libnvrtc.so: cannot open shared object file: No such file or directory . what version uf cuda should I use? mlspace environments create --env name_test --python 3.10 --cuda *
You could try vanilla Python venv instead of third party venv manager like conda or mlspace which you use. These providers could have compatibility problems with torch and cuda runtime.
Steps to reproduce
Terminal:
Code: (go through Plan)
pip3 install 'torch==2.5.0'
pip3 install para-attn
git clone https://github.com/chengzeyi/ParaAttention.git
cd ParaAttention
git submodule update --init --recursive
pip3 install 'torch==2.5.0'
pip3 install 'setuptools>=64' 'setuptools_scm>=8'
'# 'Pass --no-use-pep517 to speed up rebuild by using the legacy build system
'# 'which doesn't use a one-time tmp directory for the build
pip3 install -e '.[dev]' --no-build-isolation
'#' Or:
'# 'python3 setup.py develop
'#'Code formatting and linting
pip3 install pre-commit
pre-commit install
pre-commit run --all-files
3.torchrun --nproc_per_node=2 parallel_examples/run_cogvideox.py
What I got:
from .base import clone # noqa: E402
^^^^^^^^^^^^^^^^^^^^^^^
...
ModuleNotFoundError: No module named 'joblib'
After installing joblib i get more errors, that I can not fix. They might be related to cuda.
What I expected to get:
video in output
The text was updated successfully, but these errors were encountered: