This is your work, valued
this user is suffering from context rot
buun-llama-cpp. LLAMA Turboquant implementation with CUDA support
