leeyunseokarchive/fituna

Stop guessing your llama.cpp config. Give it a target tok/s…