-
-
Notifications
You must be signed in to change notification settings - Fork 617
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Windows GPU test is skipping #1843
Comments
Thanks for reporting @ydcjeff ! |
@vfdev-5 There seems like torch fails to find CUDA gpu or missing something in circleci config ? ..\..\tools\miniconda3\lib\site-packages\torch\cuda\__init__.py:52
c:\tools\miniconda3\lib\site-packages\torch\cuda\__init__.py:52: UserWarning: CUDA initialization: CUDA driver initialization failed, you might not have a CUDA gpu. (Triggered internally at ..\c10\cuda\CUDAFunctions.cpp:109.)
return torch._C._cuda_getDeviceCount() > 0 |
I think this is related to machine we are using on Circle CI: https://circleci.com/docs/2.0/configuration-reference/#available-windows-gpu-image which has CUDA 10.1 but we install pytorch with CUDA 11.1 support. |
I think we should do something like vision does here: https://github.com/seemethere/vision/blob/master/packaging/windows/internal/cuda_install.bat
To check the driver :
and it says
To see current processes (if connected with bash)
|
* Use cuda 10.1 on Windows Fixes #1843 * Update config.yml
🐛 Bug description
Windows GPU tests are currently not running on master
The text was updated successfully, but these errors were encountered: