Convex.jl to solve convex optimisation problems.
Convex.jl, I can use various solvers, but I cannot make the solver
verbose=false only when using Mosek.
How can I do this?
Here’s an MWE:
const SOLVER = Mosek
m = 5; n = 4
x = Variable(n, 1)
A = randn(m, n); b = randn(m, 1)
prob = minimize(sumsquares(A * x - b), [x <= 0])
prob = solve!(prob, SOLVER.Optimizer(verbose=false))
P.S. Can I copy contents in REPL to attach a result here? For reference, I’m using vim and Julia REPL now.
LOG=0, i.e. instead of
LOG=0. Note that this is a solver-specific choice, not a choice Convex made, although we should expose a generic way to suppress output (edit: here’s an issue for that: https://github.com/jump-dev/Convex.jl/issues/403).
If you update to version 0.13.7 of Convex you can try the new keyword argument
solve!(prob, SOLVER.Optimizer(), silent_solver = true)
I’ve found that
LOG=false cannot prevent other warning messages from MOSEK such as
MOSEK warning 705: #1 (nearly) zero elements are specified in sparse row ''(11) of matrix 'A'..
Can you share ideas to make them not to provide warning messages?