can't install cuda support on linux for cuda 1.1
See original GitHub issueHow to reproduce the problem
first I tried literally the code snippets from the website pip install -U spacy[cuda]
this resulted in zsh: no matches found: spacy[cuda]
Ok, my shell can’t deal with the square brackets.
Used quotation marks then to escape them, which helped here. (Might be good to add this to the docs to make copy+paste work.)
After running that…the dependent cupy build fails always:
Building wheels for collected packages: cupy
Building wheel for cupy (setup.py) ... error
ERROR: Command errored out with exit status 1:
command: /home/schmitt/git/study/explore-zendesk/.venv/bin/python -u -c 'import sys, setuptools, tokenize; sys.argv[0] = '"'"'/tmp/pip-install-cp012vja/cupy_fb58f51e771a4587abb8e187e1e8da9b/setup.py'"'"'; __file__='"'"'/tmp/pip-install-cp012vja/cupy_fb58f51e771a4587abb8e187e1e8da9b/setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(__file__);code=f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, __file__, '"'"'exec'"'"'))' bdist_wheel -d /tmp/pip-wheel-vtt8gtlq
cwd: /tmp/pip-install-cp012vja/cupy_fb58f51e771a4587abb8e187e1e8da9b/
Complete output (329 lines):
Options: {'package_name': 'cupy', 'long_description': None, 'wheel_libs': [], 'wheel_includes': [], 'no_rpath': False, 'profile': False, 'linetrace': False, 'annotate': False, 'no_cuda': False, 'use_hip': False}
-------- Configuring Module: cuda --------
-------- Configuring Module: cusolver --------
-------- Configuring Module: cudnn --------
-------- Configuring Module: nccl --------
-------- Configuring Module: nvtx --------
-------- Configuring Module: thrust --------
-------- Configuring Module: cutensor --------
/tmp/tmpmnwk_i48/a.cpp:1:10: fatal error: cutensor.h: No such file or directory
1 | #include <cutensor.h>
| ^~~~~~~~~~~~
compilation terminated.
command 'gcc' failed with exit status 1
-------- Configuring Module: cub --------
************************************************************
* CuPy Configuration Summary *
************************************************************
Build Environment:
Include directories: ['/opt/cuda/include', '/tmp/pip-install-cp012vja/cupy_fb58f51e771a4587abb8e187e1e8da9b/install/../cupy/core/include']
Library directories: ['/opt/cuda/lib64']
nvcc command : ['/opt/cuda/bin/nvcc']
Environment Variables:
CFLAGS : (none)
LDFLAGS : (none)
LIBRARY_PATH : (none)
CUDA_PATH : /opt/cuda
NVTOOLSEXT_PATH : (none)
NVCC : (none)
ROCM_HOME : (none)
Modules:
cuda : Yes (version 11010)
cusolver : Yes
cudnn : Yes (version 8005)
nccl : Yes (version 2708)
nvtx : Yes
thrust : Yes
cutensor : No
-> Include files not found: ['cutensor.h']
-> Check your CFLAGS environment variable.
cub : Yes
WARNING: Some modules could not be configured.
CuPy will be installed without these modules.
Please refer to the Installation Guide for details:
https://docs.cupy.dev/en/stable/install.html
************************************************************
... (skipped lines here)
RuntimeError: Missing file: cupy/cuda/cupy_cub.cu
Please install Cython 0.28.0 or later. Please also check the version of Cython.
See https://docs.cupy.dev/en/stable/install.html for details.
Your Environment
- Operating System: Arch Linux x86_64, 5.7.11-arch1-1
- spaCy version: 2.3.4
- Platform: Linux-5.9.11-arch2-1-x86_64-with-glibc2.32
- Python version: 3.8.6
- Environment Information: poetry with virtualenv and also tried it with conda.
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2020 NVIDIA Corporation
Built on Mon_Oct_12_20:09:46_PDT_2020
Cuda compilation tools, release 11.1, V11.1.105
Build cuda_11.1.TC455_06.29190527_0
Issue Analytics
- State:
- Created 3 years ago
- Comments:6 (4 by maintainers)
Top Results From Across the Web
NVIDIA CUDA Installation Guide for Linux
To use NVIDIA CUDA on your system, you will need the following installed: CUDA-capable GPU. A supported version of Linux with a gcc...
Read more >Install CUDA in Ubuntu, problems and workarounds
In case you can't install pytorch, make sure you are using x86_64 bit python. x86 version python is not working, there is the...
Read more >NVIDIA CUDA Installation Guide for Linux
1.1. System Requirements. To use CUDA on your system, you will need the following installed: ‣ CUDA-capable GPU. ‣ A supported version of...
Read more >Enable NVIDIA CUDA on WSL 2 - Windows - Microsoft Learn
Once you've installed the above driver, ensure you enable WSL and install a glibc-based distribution (such as Ubuntu or Debian). Ensure you have ......
Read more >Installation — pytorch_geometric documentation
Your OS. Package. CUDA. Run: PyTorch 1.13.*. PyTorch 1.12.*. Linux. Mac. Windows. Conda ... Check if PyTorch is installed with CUDA support:.
Read more >Top Related Medium Post
No results found
Top Related StackOverflow Question
No results found
Troubleshoot Live Code
Lightrun enables developers to add logs, metrics and snapshots to live code - no restarts or redeploys required.
Start FreeTop Related Reddit Thread
No results found
Top Related Hackernoon Post
No results found
Top Related Tweet
No results found
Top Related Dev.to Post
No results found
Top Related Hashnode Post
No results found
Top GitHub Comments
Right, spacy v2.3.4 (really thinc
<=7.4.3
) is not compatible with cupy v8.These are the only releases on pypi for
cupy-cuda111
: https://pypi.org/project/cupy-cuda111/#historyThis thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.