Add Support for More Local LLM Inferencing Tools

[jmemcc]

Issue

Setup currently supports only Ollama for local LLM inferencing. Expanding support to other local LLM inferencing tools (e.g. vLLM, llama.cpp, koboldcpp) will allow users to work with the inference tooling they are most comfortable with, or best suits their needs.

Solution

Add support for other inference tools into the chat interface, and alter the Docker setup to allow the inference tools to be run inside the Docker container, not just via a bridge to the host's machine.