How to pass optimizer options to GalacticOptim?

Specific Domains Optimization (Mathematical)
1

I’m trying to play with GalacticOptim. How can I pass options to the optimizer? In plain Optim, I can pass options like store_trace with Optim.Options(store_trace=true). How can I pass those options in the example below?

using GalacticOptim
using Optim

function rosenbrock_n(x)
    n1 = length(x) - 1
    sum(100(x[i+1] - x[i]^2)^2 + (1 - x[i])^2 for i in 1:n1)
end

n = 2
x0 = 10 .* ones(n)

rosenbrock_ga(x, _) = rosenbrock_n(x)
go_fun = OptimizationFunction(rosenbrock_ga, GalacticOptim.AutoFiniteDiff())
go_obj = OptimizationProblem(go_fun, x0)
go = GalacticOptim.solve(go_obj, GradientDescent())
2

It should just extend.

go = GalacticOptim.solve(go_obj, GradientDescent(), show_trace = true)
2 Likes