Add support for gguf

[phronmophobic]

The latest llama.cpp development has deprecated the ggml format in favor of a new gguf format.

llama.cpp has chosen to break their API and make ggml models useless. The goal for llama.clj is to upgrade without breaking backwards compatibility. More research is required, but the initial plan is something like:

• treat the new raw API for llama.cpp as a separate library
• create a protocol for any shared functionality and implement it for both the ggml version and the latest version
• create independent builds that can be included independently or together
• Add better support for including your own llama.cpp build

[phronmophobic]

Links:

• GGUF (Breaking Change to Model Files) abetlen/llama-cpp-python#633
• Add Support for the new GGUF format which replaces GGML oobabooga/text-generation-webui#3676
• GGUF Support abetlen/llama-cpp-python#628
• https://github.com/philpax/ggml/blob/gguf-spec/docs/gguf.md
• llamma.cpp breaking change deprecating GGML in favor of GGUF ollama/ollama#423
• GGUF ggerganov/llama.cpp#2398
• https://www.reddit.com/r/LocalLLaMA/comments/15xycn2/llamacpp_gguf_merged/

[phronmophobic]

I updated to the latest version of llama.cpp locally and was able to get a gguf model to run without too many changes. However, there are a still a few updates in progress for llama.cpp that I'll probably wait on before making a new release:

• metal support not currently working: Metal uses Intel HD instead of AMD GPU on Intel iMac5k  ggerganov/llama.cpp#2407
• new decoding API: llama : custom attention mask + parallel decoding + no context swaps ggerganov/llama.cpp#3228

[phronmophobic]

Fixed in v0.8.