Parametric meta #12
Parametric meta #12
Conversation
|
@tmigot I think I registered here : JuliaSmoothOptimizers/NLPModels.jl@e3b8b73 , will this fix your tests? |
Great, thank you! I will wait for the official release and then rebase. |
Codecov Report
@@ Coverage Diff @@
## master #12 +/- ##
==========================================
+ Coverage 97.92% 98.03% +0.11%
==========================================
Files 6 6
Lines 578 612 +34
==========================================
+ Hits 566 600 +34
Misses 12 12
Continue to review full report at Codecov.
|
|
|
src/feasibility-form-nls.jl
Outdated
| x0 = [meta.x0; zeros(nequ)], | ||
| lvar = [meta.lvar; fill(-Inf, nequ)], | ||
| uvar = [meta.uvar; fill(Inf, nequ)], | ||
| x0 = [meta.x0; zeros(T, nequ)], |
There was a problem hiding this comment.
I think this fails for GPU but I don't see a viable alternative, do you?
There was a problem hiding this comment.
Would this work?
x0= zero(meta.x0, meta.nvar + nequ)
x0[1 : meta.nvar] .= meta.x0There was a problem hiding this comment.
I didn't test it, but this tutorial suggests that cat and fill are available methods for GPUArrays.
|
|
|
@dpo I updated the constructors. What do you think? |
|
|
Thanks. And Probot auto-merged, but it was in accordance with what we discussed yesterday. |
4 participants
Follows NLPModels.jl PR #353 with parametric meta.
Add tests on SimpleNLPModel and SimpleNLSModel for Float64 and Float32.