AI tools unable to find NVIDIA drivers

Support Graphics & Display
1

Hi all,

I’ve been wanting to try a couple of AI-related Github projects recently, but without success. I believe it has to do with the Python environments being unable to work with my GPU.

Example of an error I got when attempting to launch a tool called Fooocus:

Traceback (most recent call last):
  File "/home/.../Documents/programming/Fooocus/launch.py", line 152, in <module>
    from webui import *
  File "/home/.../Documents/programming/Fooocus/webui.py", line 10, in <module>
    import modules.async_worker as worker
  File "/home/.../Documents/programming/Fooocus/modules/async_worker.py", line 3, in <module>
    from extras.inpaint_mask import generate_mask_from_image, SAMOptions
  File "/home/.../Documents/programming/Fooocus/extras/inpaint_mask.py", line 6, in <module>
    from extras.GroundingDINO.util.inference import default_groundingdino
  File "/home/.../Documents/programming/Fooocus/extras/GroundingDINO/util/inference.py", line 3, in <module>
    import ldm_patched.modules.model_management as model_management
  File "/home/.../Documents/programming/Fooocus/ldm_patched/modules/model_management.py", line 121, in <module>
    total_vram = get_total_memory(get_torch_device()) / (1024 * 1024)
  File "/home/.../Documents/programming/Fooocus/ldm_patched/modules/model_management.py", line 90, in get_torch_device
    return torch.device(torch.cuda.current_device())
  File "/home/.../.virtualenvs/fooocus/lib/python3.10/site-packages/torch/cuda/__init__.py", line 1069, in current_device
    _lazy_init()
  File "/home/.../.virtualenvs/fooocus/lib/python3.10/site-packages/torch/cuda/__init__.py", line 410, in _lazy_init
    torch._C._cuda_init()
RuntimeError: Found no NVIDIA driver on your system. Please check that you have an NVIDIA GPU and installed a driver from http://www.nvidia.com/Download/index.aspx

Overview of my System and Graphics card:

$ inxi -Fza
System:
  Kernel: 6.12.61-1-MANJARO arch: x86_64 bits: 64 compiler: gcc v: 15.2.1
    clocksource: tsc avail: hpet,acpi_pm
    parameters: BOOT_IMAGE=/boot/vmlinuz-6.12-x86_64
    root=UUID=787d3c90-81c0-4833-bcb0-008c4e49bf60 rw quiet splash apparmor=1
    security=apparmor udev.log_priority=3
  Desktop: Xfce v: 4.20.1 tk: Gtk v: 3.24.51 wm: xfwm4 v: 4.20.0
    with: xfce4-panel tools: xfce4-screensaver vt: 7 dm: LightDM v: 1.32.0
    Distro: Manjaro base: Arch Linux
..
Graphics:
  Device-1: NVIDIA GB205 [GeForce RTX 5070] vendor: ASUSTeK driver: N/A
    alternate: nouveau non-free: 550-580.xx+ status: current (as of 2025-11)
    arch: Lovelace code: AD1xx process: TSMC n4 (5nm) built: 2022+ pcie:
    gen: 3 speed: 8 GT/s lanes: 16 link-max: gen: 5 speed: 32 GT/s
    bus-ID: 01:00.0 chip-ID: 10de:2f04 class-ID: 0300
  Display: x11 server: X.org v: 1.21.1.21 compositor: xfwm4 v: 4.20.0
    driver: X: loaded: modesetting,nouveau alternate: fbdev,nv,vesa gpu: N/A
    display-ID: :0.0 note: <missing: xdpyinfo/xrandr>
  Monitor-1: Unknown-1 size-res: N/A modes: 1920x1080
  API: EGL v: 1.5 platforms: device: 0 drv: swrast surfaceless: drv: swrast
    x11: drv: swrast inactive: gbm,wayland
  API: OpenGL v: 4.5 vendor: mesa v: 25.3.1-arch1.2 glx-v: 1.4
    direct-render: yes renderer: llvmpipe (LLVM 21.1.6 256 bits)
    device-ID: ffffffff:ffffffff memory: 15.18 GiB unified: yes
  Info: Tools: api: eglinfo,glxinfo de: xfce4-display-settings x11: xprop

Overview of drivers:

$ mhwd -l
> 0000:01:00.0 (0300:10de:2f04) Display controller nVidia Corporation:
--------------------------------------------------------------------------------
                  NAME               VERSION          FREEDRIVER           TYPE
--------------------------------------------------------------------------------
          video-nvidia            2025.09.29               false            PCI
    video-nvidia-575xx            2025.09.29               false            PCI
    video-nvidia-570xx            2025.09.29               false            PCI
           video-linux            2024.05.06                true            PCI
     video-modesetting            2020.01.13                true            PCI
            video-vesa            2017.03.12                true            PCI


$ mhwd -li                                                                           
> Installed PCI configs:
--------------------------------------------------------------------------------
                  NAME               VERSION          FREEDRIVER           TYPE
--------------------------------------------------------------------------------
           video-linux            2024.05.06                true            PCI


Warning: No installed USB configs!

I have no idea how to troubleshoot this. Any suggestions other than checking the README and issue list of this specific project? I have experienced similar issues with other AI-projects, so I assume my set-up is to blame.

2

Install the drivers

sudo mhwd -i pci video-nvidia

If your system is a dual-gpu laptop install the hybrid driver instead (either intel or amd)

sudo mhwd -i pci video-hybrid-{intel or amd}-nvidia-prime
3 Likes
3

I’m not familiar with the term dual-gpu. Does it refer to a Windows+Manjaro dual boot? That’s indeed what I’m currently working with.

I tried your first suggestion first, despite assuming video-nvidia is already installed as stated in my OP. That didn’t do the trick.

Then I tried your second suggestion, going with the intel variant. It gave me a warning

Warning: no matching device for config 'video-hybrid-intel-nvidia-prime' found!

but then proceeded to install it anyway. It gives me the same warning when I go for the amd variant, (but I’m pretty sure the intel variant should be correct).

When I try to launch the tool now (with video-hybrid-intel-nvidia-prime installed), I’m getting a new error:

RuntimeError: No CUDA GPUs are available

So that feels like progress, but may very well not be. Any suggestions?

4

It refers to a hybrid where graphics may be provided by either an integrated GPU (iGPU) or a discrete GPU (dGPU) thus having two distinct graphics processors.

If your system is a desktop then it does not matter as only laptops use the hybrid driver.

The mhwd -li you provided clearly shows the video-nvidia driver as not installed.

sudo mhwd -i pci video-nvidia

Please provide the full output from inxi -Fxxxyc0

2 Likes
5

Oh okay, I was completely misunderstanding the commands that I posted. And for some reason I thought I had installed video-nvidia during this process, but apparently I didn’t. It looks like this fixed the problem for me. Thanks!

P.S the tool is still not working, but that’s outside of the scope of this thread.

6

This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.