Add github protocol pulling and http://

[ericcurtin]

As pulling protocols to llama-run

[ggerganov]

Btw, it would be great QoL if the llama-run command does not re-download a model when using a remote link. Do you plan to implement this?

[ericcurtin]

Btw, it would be great QoL if the llama-run command does not re-download a model when using a remote link. Do you plan to implement this?

We added this recently on line 326:

if (std::filesystem::exists(output_file))

which should do this, unless you found a bug @ggerganov 😅

[ggerganov]

It works now, thanks.