i have a server running Ubuntu with two Xeon x5680s and just install a NVIDIA Tesla k40c HPC / GPU for running PyTorch and tensorflow-gpu models but cant get the drivers running. i used "sudo ubuntu-drivers autoinstall" which tried to install "nvidia-driver-470" but aborted when trying to install nvidia-cudnn in the process:
Removing nvidia-cudnn (8.2.4.15~cuda11.4) ...
Purging cuDNN installation from /usr
dpkg: error processing package nvidia-cudnn (--remove):
installed nvidia-cudnn package pre-removal script subprocess returned error exit status 1
dpkg: too many errors, stopping
Errors were encountered while processing:
nvidia-cudnn
Processing was halted because there were too many errors.
E: Sub-process /usr/bin/dpkg returned an error code (1)
How do i get my gpu running nvidia drivers it also install nvidia-smi but when i run that, it throws this error:
NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.
obviously trying to install "tensorflow-gpu" via pip it throws an error and PyTorch can also not use the gpu.
