Hacker News Reader

llama.cpp sycl and vllm xmx work is pretty incredible right now - you just gotta build it with some extra flags

Llama would be nice for the ggufs. Any specific flags or tutorials I should look at?

The docs are a great start.