local-apps: update node-llama-cpp snippet

[giladgd]

This PR updates the code snippet of node-llama-cpp to use a tag-based model URI, like in llama.cpp and Ollama.
The implementation for this uses the Ollama support on Hugging Face, and is based on this PR: ggml-org/llama.cpp#11195

Examples:

# chat with the model
npx -y node-llama-cpp chat hf:mradermacher/Llama-3.2-3B-Instruct-GGUF:Q4_K_M

# estimate the model compatibility with the current machine hardware
npx -y node-llama-cpp inspect estimate hf:mradermacher/Llama-3.2-3B-Instruct-GGUF:Q4_K_M

[Vaibhavs10]

Amazing! Let's wait for @ngxson to give an LGTM too and merge!

[julien-c]

nice, this looks quite clean 👍

[giladgd]

@ngxson Let me know if you need me to make changes to the PR to merge it

[giladgd]

Do you need any input from me to merge this PR? It'd be a bummer to see this PR get stale...

[julien-c]

@ngxson i think we can merge if you approve

[ngxson]

The failed CI doesn't relate to this PR, so I assume it's ok to merge?

[julien-c]

The failed CI doesn't relate to this PR, so I assume it's ok to merge?

yes imo