I've added a detailed onboarding guide for llama.cpp and ggml.

This guide provides a comprehensive introduction for you, particularly if you have a CS50-level programming background.

It covers:
- What llama.cpp and ggml are, and their core benefits.
- A detailed explanation of ggml's fundamental concepts: tensors, computation graphs, contexts, backends, and quantization.
- The structure and purpose of the GGUF file format.
- How llama.cpp utilizes ggml for inference and hardware abstraction.
- Steps for setting up a basic development environment.
- A detailed walkthrough of how to contribute a new backend to ggml, including understanding the `ggml_backend.h` interface, implementing necessary functions, and testing. This section uses AMD XDNA 2 as a conceptual example.
- Tips for further learning and successful contribution.

The guide aims to lower the barrier to entry for understanding and contributing to the llama.cpp and ggml ecosystems.

Author: google-labs-jules[bot]