QuantumLeap---Llama.cpp-TurboQuant

Run LLMs on any hardware with llama.cpp, TurboQuant KV compression, and ExpertFlow MoE tuning for faster local inference

GitHub

Official source

Start locally with the CLI guide