So i’ve set up manjaro 6.1.41-1. I wanted to use tensorflow accelerated by GPU for which i need CUDA, cuDNN and bazel but can’t find none in AUR repository that fits 2.12.0 tensorflow, I tried to find another tensorflow and its dependencies but couldn’t find anything too.
Welcome to the forum!
tensorflow-cuda 2.13.0-1 in all branches. Are you fully up to date? See Pacman-mirrors - Manjaro
I still get
tensorflow/tsl/cuda/cudart_stub.cc:28] Could not find cuda drivers on your machine, GPU will not be used.
The repo versions are 2.13.0.
If you want older versions, you have to setup a python virtualenv and install it there via pip.
So few days later, still can’t make it run on Manjaro.
NVIDIA-SMI 535.86.05 Driver Version: 535.86.05 CUDA Version: 12.2 |
nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2023 NVIDIA Corporation Built on Tue_Jun_13_19:16:58_PDT_2023 Cuda compilation tools, release 12.2, V12.2.91 Build cuda_12.2.r12.2/compiler.32965470_0
at this point i think its impossible to make tensorflow run on GPU on Manjaro
Can’t find any usable resources that would help with this problem. I still get following error:
I tensorflow/tsl/cuda/cudart_stub.cc:28] Could not find cuda drivers on your machine, GPU will not be used.
How did you install? The pip version is compiled against Cuda 11.8 but you’re using 12.2.
I installed tensorflow-opt-cuda throught pacman. Oh and the cuda listed in smi are proprietary ones