Feat all quants ci

[nguyenhoangthuan99]

Fix janhq/cortex.cpp#1155
Summary changes:

• update model yml script to download model.yml in cortexso hugging face, then update any field, keep all comments from original file.
• CI pipeline for update model.yml, can input quantization level, model size, and model repo_id
• CI pipeline for convert and quantize model from huggingface format to gguf format. Can input quantization level, model size, and model repo_id

[vansangpfiev]

Should use the same name for update model? I see we have update-model-yml.yml and update_model_yaml.py

[dan-homebrew]

@nguyenhoangthuan99 Thanks. I've given a preliminary approval to unblock.

[dan-homebrew]

@hiento09 @nguyenhoangthuan99 For converter and quantization, would it make sense for us to pull it out into a separate repo?

• llama.cpp engine is meant to be for engines
• I see our CI running for to build this for every Hardware and OS, when it doesn't really make sense