Add support for GGUF machine learning model files

[fegge]

This PR adds support for the GGUF file format. GGUF is used to store machine learning models for executors based on the GGML machine learning library like llama.cpp and whisper.cpp developed by Meta.

Since there was no obvious location to add machine-learning model file formats, and since there are multiple other ML specific files formats that could be added in the future, I chose to create a new top-level ml directory and added the gguf.ksy file under that directory. I'm happy to put it somewhere else if someone has opinions on the location.

Example files for testing can be found under the models directory in the llama.cpp repository. (Small model files can be generated using the gguf utility in llama.cpp.)