Implement Decentralized RAG based system for supporting Distributed LLM models

[Super-Genius]

Resources on this

https://docs.gnus.ai/gnus.ai-gitbook/technical-information/ai-systems/overview

ChatLLM - RAG based LLM chat module - https://github.com/foldl/chatllm.cpp
GGLM - code for multiple shaders/models for LLM - https://github.com/ggerganov/ggml
Vulkan Shaders to load into MNN - https://github.com/ggerganov/ggml/tree/master/src/vulkan-shaders

We really need a small JSON meta language to load and process data and aggregate it for this.

Then an Chat APP can use the distibuted LLM RAG models like the sample code.