-
-
Notifications
You must be signed in to change notification settings - Fork 183
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
Expose variadic arguments for algebra_solver #2818
Comments
The variadic interface would completely subsume the original, correct? I only ask because we currently have a technical restriction in Stanc where a variadic function cannot have non-variadic overloads |
My plan is to only expose variadic arguments for But no, no overloading for variadic functions. |
I don’t think my question was clear, we currently already have an exposed function called |
Thanks for clarifying. Yes, you're right. We actually won't break backwards compatibility. I got a little bit mixed up: in stan-math we do break backwards compatibility because we have to reposition the |
|
Description
The algebraic solvers,
algebra_solver_newton
andalgebra_solver_powell
aren't using variadic arguments. Most of the code seems to be already there but it is all wrapped in a function that still uses the strict function signature. We need to create the following updated signatures:algebra_solver_newton(f, x0, ...)
algebra_solver_tol_newton(f, x0, scaling_step_size, function_tolerance, max_num_steps, ...)
algebra_solver_powell(f, x0, ...)
algebra_solver_tol_powell(f, x0, relative_tolerance, function_tolerance, max_num_steps, ...)
Open question: should we keep the
relative_tolerance
tuning parameter used for Powell's method? Right now, the two solvers have slightly different choices of tuning parameters, which has to do with the underlying algorithms they use. I think that's fine but it might be confusing to the user.Current Version:
v4.4.0
The text was updated successfully, but these errors were encountered: