Fix input for (L)BFGS and PSO #431
Conversation
| @inbounds state.x[i] = state.x[i] + state.dx[i] | ||
| @simd for i = 1:n | ||
| state.dx[i] .= state.alpha .* state.s[i] | ||
| state.x[i] .= state.x[i] .+ state.dx[i] |
There was a problem hiding this comment.
Will it not work to do state.x .= state.x .+ state.dx? (Would dot-notation not work for arbitrary vector-types?)
There was a problem hiding this comment.
Sorry, I got confused. I mean on the state.s: I thought that's meant to be a vector, whilst state.x may be a multidimensional array?
There was a problem hiding this comment.
This is a mistake! I had originally tried to change it to broadcast, but that didn't like mixing vectors and matrices.
I mean, state.x is allowed to be multidimensional. If x is passed as a matrix we want Optim.minimizer(res) to be a matrix as well.
We could store s as the same type as x, and throw vec's around state.s whereever it shows up?
There was a problem hiding this comment.
Would reshape on state.s work?
Alternatively reshape on Optim.minimizer(res), and then just have state.x as a vector internally
There was a problem hiding this comment.
yeah, ofc you could just do minimizer = similar(initial_x); copy!(minimizer, state.x) in the end, that's another possibility.
Codecov Report
@@ Coverage Diff @@
## master #431 +/- ##
==========================================
+ Coverage 90.46% 90.84% +0.38%
==========================================
Files 32 32
Lines 1562 1562
==========================================
+ Hits 1413 1419 +6
+ Misses 149 143 -6
Continue to review full report at Codecov.
|
test/multivariate/array.jl
Outdated
| @test typeof(_minimizer) <: Array{Float64, 3} | ||
| @test size(_minimizer) == (2,2,1) | ||
| if !(m in (SimulatedAnnealing, ParticleSwarm)) | ||
| println(typeof(m)) |
There was a problem hiding this comment.
yes yes, this is debugging text, will remove... :)
ref #399