Replies: 1 comment
-
|
The scriberr-cuda image released is only for amd64.. I did not release a version for arm. I simply assumed cuda users will all be use x86. My bad.. First time I'm seeing interest in using arm based cuda devices.. Very curious to see how the performance is. Regarding the out of memory error, it is almost definitely a permission error.. If you are using a bind mount, can you check the ownership of the scriberr data and env folders and verify it matches the PUID/PGID you specified in compose? To be safe, can you also |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
Hi there, I tried to install Scriberr on my Nvidia Orin Nano which has an ARM64 architecture and an Nvidia CUDA. So I used the docker image as specified in the Readme but I got an error message "scriberr The requested image's platform (linux/amd64) does not match the detected host platform (linux/arm64/v8) and no specific platform was requested".
As I saw somewhere in Reddit that Scriberr could run on a Raspberry Pi, I thought the specified image could run on my Orin.
How do you specify the ARM64 target ?
Having failed I then tried to install Scriberr on my Windows PC. So I started with a fresh WSL Ubuntu 24, installed docker and the Nvidia Container Toolkit and tried to run the same docker command, but there I got the "out of memory (14)" error despite having set the proper PUID/PGID.
Beta Was this translation helpful? Give feedback.
All reactions