Electric Eel allowed a tickbox, when installing an app, to load “the Nvidia drivers” (or when you installed the app catalogues, or something).
Does that mean that I can run a docker compose which wants to use nvidia drivers, and it should work?
Or do I need to separately install the Nvidia toolkit? I hope not - it looks too difficult for me in TrueNAS and it might wreck it!
I did try to follow earlier threads as best I could but I didn’t understand.
This is what I want to do in my docker compose:
services:
frigate:
...
image: ghcr.io/blakeblackshear/frigate:stable
deploy: # <------------- Add this section
resources:
reservations:
devices:
- driver: nvidia
device_ids: ['0'] # this is only needed when using multiple GPUs
count: 1 # number of GPUs
capabilities: [gpu]
having tried it, Frigate seems to crash out after a few minutes (and I don’t yet know why but I presume it might be Nvidia related).
I forgot to add: running the Frigate app and it uses Nvidia OK and indeed its shell shows me this:
That’s useful to know - thanks. I’ll try some more experiments, based on the assumption that Nvidia drivers are effectively an integrated part of the TrueNAS appliance and not in need of additional work by me.
edit: I get an error Error response from daemon: could not select device driver "nvidia" with capabilities: [[gpu]]
even though my docker compose says the same as yours:
I did some more searching (a lot more!) and found the AI help system in Frigate docs quite useful. The problem I describe above is due to me not having the Nvidia toolkit installed, because I thought having access to the Nvidia drivers would suffice. I was wrong: I have opened a separate post and I refer to it here