Open WMCh opened 4 years ago
libOpenGL.so
is missing. Thus, your nvidia driver/opengl installation is broken. You need to reinstall the driver, or see if libOpenGL.so
has perhaps been put in a non-standard location (just searching the whole system for *GL*
is simplest). What linux distro are you using?
libOpenGL.so
is missing. Thus, your nvidia driver/opengl installation is broken. You need to reinstall the driver, or see iflibOpenGL.so
has perhaps been put in a non-standard location (just searching the whole system for*GL*
is simplest). What linux distro are you using?
I appreciate your reply. I've searched the whole system for the libOpenGL.so
and it turns out to be missing. However, the command nvidia-smi
works (it shows my driver version is 418.56) and other tensorflow codes run well. Does it mean that my nvidia driver is OK? I installed the OpenGL with the apt-get
command as the Internet told me. In addition, my linux distro is Ubuntu 16.04.
libOpenGL.so
is included in the nvidia driver installation, so something has gone wrong, even if cuda stuff works. It is quite easy to 'lose' the GL libraries as they can be overwritten by other packages like mesa (installing OpenGL/etc. through apt
can actually make the situation worse, as this can overwrite the nvidia version). If you can reinstall the driver, that should fix the problem.
nsorflow codes run well. Does it
have you solved the problem?
nsorflow codes run well. Does it
have you solved the problem?
It seemed that my NVIDIA driver was broken as @pmh47 said. The problem was fixed when I changed another docker image. Maybe you can try to reinstall the driver.
Hi, when I run command
pip install .
, the error information is as below:I have referenced to other issues and made some effort, but it doesn't work. The output of
ls -l /usr/lib/*/*GL*
is:Is there any advice? Thank you very much!