Only warn of rate-limits when using HF endpoint

[HennerM]

I am trying the llm-vscode extension with llm-ls on a locally hosted endpoint (running a custom fine-tuned model), however the extension still gives a warning that I might get rate limited by HuggingFace.

Since inference doesn't run on a HuggingFace server this warning is not necessary.

[McPatate]

Since we know have the adaptor setting, I'd rather check its value than doing so w/ the URL. Wdyt?

[HennerM]

Since we know have the adaptor setting, I'd rather check its value than doing so w/ the URL. Wdyt?

Yes good idea, I changed this to just check the adopter value now

[McPatate]

Small nits on naming, otherwise we should be good to go!