I have a brm model which I would like to use it for predictions in a system in my company. However, installing brms (actually rstan) is not an option.
Is there a way to make predictions using the model without loading the brms?
If you just do brms::posterior_predict(...) it won’t actually load brms, just call the posterior_predict() function from brms, but I’m not sure if that’s what you’re looking for.
Is RStan’s GPL-3 license the problem? If so then you can use our new CmdStanR interface as the backend for brms instead of RStan (CmdStanR has a much more permissive BSD license).