Being able to use local Llama.cpp models

[Magnarks]

Currently, if I want to use the Llama.cpp provider, it only lets me select certain models; it doesn't allow me to use the ones I have downloaded, unlike Ollama, which has the function of automatically detecting models.