Removing LLAMA from the AUR's ggml-vulkan build

[mohamadali-halwani]

Greetings everyone
I hope this discussion finds you well,

I am a medical practitioner and I'm new to the opensource/Build and compile/superuser thing.
I Don't like LLAMA and I would like to replace or remove it from the build. I prefer using something like qwen, deepseek, or mistral. Can you suggest any tutorials/courses that teach me how to do that? Or is there any documentation I should read to learn how to achieve this?

If it's not possible to edit a program from the AUR, should I fork this repo and try to hard code it instead, or what should I do?

Any fruitful suggestion is welcome, and I'm happy to be in this community.

Best Regards,
MohamadAli Halwani

[Green-Sky]

1. This is the ggml repository. It itself has not much "llama" specific code.
2. The llama.cpp project outgrew its initial focus on just llama rather quickly. It supports qwen, deepseek, mistral and many more. So the name "llama.cpp" is just a remnant of what once was. There is no reason to "remove" any code related to a specific model, also most models share some parts of the architecture.
3. For AUR specific questions you would have to talk to the packager(s). It is a common misconception that the creators of a project are also doing the package maintaining for distribuitons. (I don't know who that is and that is out of scope for this question anyway)

[mohamadali-halwani]

Thanks a lot <3