Add --fit, --parallel options and improve GGUF parser

Update version follow PR #5 merge

Merge pull request #5 from KeinNiemand/main

Merge branch 'main' of https://github.com/thad0ctor/llama-server-launcher

Allow --cpu-moe option to work with ik_llama

fix flash attention argument for newest llama.cpp api

GPU layer override

Reverted FA launch parameters

updated launching for llama.cpp changes to flash-attn enabling

Fix mmproj text formatting and version #

Add checkbox control for mmproj scanning performance

add mmproj cmd

Added MOE options - llama.cpp PR 15077

Added Q6_K kv cache option

fix config issues