-
-
Notifications
You must be signed in to change notification settings - Fork 67
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
cmdstanpy cmdstanr name consistency for sampler #221
Comments
@jgabry - willing to go with CmdStanR versions - comments? |
Are both going to be consistent with CmdStan itself? The only reason someone's suggesting
|
The reason we've been using |
@jgabry or @mitzimorris : can one of you clarify the issue to reflect this decision? The current text implies there'a a decision to make. Just keep in mind that whatever gets decided here is probably going to have to carry over to a simplified CmdStan3. My guess is that most Stan users migrating to CmdStanR or CmdStanPy will be coming from RStan or PyStan. Anyway, not a big deal. I don't want to derail development with a long digression on naming. |
Basically the decision is when naming arguments should we: a) Mirror current cmdstan names when possible |
I’ve been doing (a) because that seemed simplest/cleanest until there is CmdStan3. When that happens we can then do whatever CmdStan3 does. |
I'd strongly prefer whatever decision is made now for CmdStanPy and CmdStanR to be carried forward for CmdStan3. I think that's more important than the decision itself. |
In that case we should do (c), since I don't think either (a) or (b) will result in the most sensible names for CmdStan3 to use. @mitzimorris I think that means we should have another meeting and get on the same page for all the names |
this issue was filed by mistake - after some discussion with folks in the office and checking against RStan and CmdStan, I filed the 2nd issue on CmdStanR, because I think that "chains" and "cores" is better. |
What about the other arguments like |
I started off thinking c) was the best option, hence 'warmup_iters', 'sampling_iters', and 'step_size'. but then folks raised objections about confusion coming from existing interfaces and all of the teaching materials and stack overflow, forums, etc discussions which are what is most likely to be pulled back by a search engine.
keeping issue open - changed title and description |
Summary:
improve consistency of argument names between CmdStanPy and CmdStanR
Description:
different names to sampler function
after discussion here: stan-dev/design-docs#15, argument names for sampler methods are:
cores
(current name)chains
(current name)iter_warmup
instead of "warmup_iters"iter_sampling
instead of "sampling_iters"Additional Information:
Provide any additional information here.
Current Version:
The text was updated successfully, but these errors were encountered: