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
I have installed CUDA and was about to build OpenCV 4. In doing so, I noticed a whole series of build errors. The reason is probably the used CUDA version 12.4. Unfortunately it is quite difficult to find out which CUDA version is currently working. As of OpenCV 4.9, CUDA 12.2 is probably supported. Wouldn't it make sense to throw an error message in CMake when recognizing CUDA if the recognized CUDA version is not supported? Or to deactivate the CUDA parts?
Additional context
No response
The text was updated successfully, but these errors were encountered:
You would also need to check the version of cuDNN and the Nvidia Video Codec SDK. As of 18/03/24 the latest version of all three are not compatible with OpenCV 4.9.0 (released 28/12/2023) see
Describe the feature and motivation
I have installed CUDA and was about to build OpenCV 4. In doing so, I noticed a whole series of build errors. The reason is probably the used CUDA version 12.4. Unfortunately it is quite difficult to find out which CUDA version is currently working. As of OpenCV 4.9, CUDA 12.2 is probably supported. Wouldn't it make sense to throw an error message in CMake when recognizing CUDA if the recognized CUDA version is not supported? Or to deactivate the CUDA parts?
Additional context
No response
The text was updated successfully, but these errors were encountered: