[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#1063673: ITP: llama.cpp -- Inference of Meta's LLaMA model (and others) in pure C/C++



On 2025-02-06 09:26, Petter Reinholdtsen wrote:
> I hope so too, but I guess we will soon find out.  My initial draft on
> <URL: https://salsa.debian.org/deeplearning-team/whisper.cpp > will need
> a lot of updates  to bring it in line with this new approach. :)

If it helps, after finally achieving a satisfying result after many
iterations, I deliberately re-wrote the history of the llama.cpp repo to
(1) start with a very minimal packaging for the baseline, and then to
(2) add feature by feature.

In fact, since practically all of the backend configuration/performance
stuff is actually just about ggml, which llama.cpp and whisper.cpp
share, you should be able to mirror every commit after the first.

Happy to help, let me know if I should do anything.

Best,
Christian


Reply to: