So I am playing with llama.cpp, ollama, ComfyUI and I cannot figure out how to tell them to share the local models that I am downloading. Is there a local LLM "standard" for where to store these files?