AVX512 Build

[sfxworks]

I've been having trouble building localAI to work with avx512. According to your makefile, even with setting rebuild to true with docker, it only builds support for avx2. I'd like to have avx512 support for my machine so I can run models faster.

[mudler]

@sfxworks can you try master builds/container images and share logs with --debug enabled ? it should now support avx512, thank you!

[sfxworks]

With the latest container build, I still see it loading with avx2. This was on my avx512 machine.
�[90m7:10AM�[0m �[32mINF�[0m �[1m[llama-cpp] attempting to load with AVX2 variant�[0m

Here is a full log from container build to chatting with an AI model in the UI.

logs.txt