LLM Examples: confusing flow for Python and C++

[bconsolvo]

Starting with the overview (https://ryzenai.docs.amd.com/en/latest/llm/overview.html), I do find that the documentation is confusing, pointing to a lot of places. Can you make this a 1-click / 1 command line to get set up, and detect the HW along the way?

To get where I needed, I followed this flow:

1. Start with Ryzen AI LLM overview: https://ryzenai.docs.amd.com/en/latest/llm/overview.html
2. HF hybrid models: https://huggingface.co/collections/amd/ryzenai-15-llm-hybrid-models-6859a64b421b5c27e1e53899
3. Find a model card: https://huggingface.co/amd/Phi-3-mini-4k-instruct-awq-g128-int4-asym-fp16-onnx-hybrid; then it points to OGA docs
4. OnnxRuntime GenAI (OGA) Flow: https://ryzenai.docs.amd.com/en/latest/hybrid_oga.html. Need to install:
   • GPU driver
   • Git for Windows

Notes: Confusing on whether I can just start from "C++ Program" or "Python Script" in documentation. I just skipped to "Python Script" but found out I needed to go back into the C++ Program. Do I just need to install the model from HF first, or do I need to copy all of the required DLL / txt / exe files?

Would rather not have to git clone the whole HF repo. Just would rather point to the model on HF and use the HF API.