Re: Ollama package ships no CUDA backend, NVIDIA users left with slower Vulkan path

Hi!

On 22.07.26 um 16:42 Neal Gompa wrote:
> RPM Fusion is not part of the Fedora Project. You'll need to talk to
> them if you want ollama-cuda. From the Fedora point of view, I would
> hope that we'd push harder for everything to use Vulkan Compute or
> OpenCL instead of these disparate proprietary APIs. And while ROCm is
> open source, it is a vendor-specific API, and I would love for us to
> not rely on that in the future.

Are there any considerations about making pytorch using OpenCL?
I only know about this half done project:
https://github.com/artyom-beilis/pytorch_dlprim

Best regards
Christoph

-- 
_______________________________________________
devel mailing list -- [email protected]
To unsubscribe send an email to [email protected]
Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: https://lists.fedoraproject.org/archives/list/[email protected]
Do not reply to spam, report it: https://forge.fedoraproject.org/infra/tickets/issues/new