-
Notifications
You must be signed in to change notification settings - Fork 34
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
* Simplify testset loop
- Loading branch information
Showing
3 changed files
with
141 additions
and
148 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -1,3 +1,3 @@ | ||
| OptimTestProblems | ||
| #DoubleFloats | ||
| DoubleFloats 0.1.10 | ||
| Optim 0.16 |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -1,148 +1,141 @@ | ||
| #doublefloattypes = [typeof(Double64(1)), typeof(Double32(1))] | ||
| # DoubleFloats breaks because of | ||
| # https://github.com/JuliaMath/DoubleFloats.jl/issues/18 | ||
| # Include them again when (if) it is fixed. | ||
| doublefloattypes = [] | ||
|
|
||
| for T in [Float32, Float64, BigFloat, | ||
| doublefloattypes...] | ||
| @eval begin | ||
| @testset "Arbitrary precision - initial step guess: $($T)" begin | ||
| f(x) = dot(x, x) | ||
| function g!(out, x) | ||
| out[:] = 2x | ||
| end | ||
|
|
||
| x = convert.($T, [-1, -1]) | ||
| df = NLSolversBase.OnceDifferentiable(f,g!,x) | ||
|
|
||
| phi_0, grtmp = NLSolversBase.value_gradient!(df, x) | ||
| @test !isnan(phi_0) | ||
| @test phi_0 isa $T | ||
| @test !any(isnan, grtmp) | ||
| @test grtmp isa Vector{$T} | ||
| p = -grtmp | ||
| dphi_0 = dot(p, grtmp) | ||
| @test !isnan(dphi_0) | ||
| @test dphi_0 isa $T | ||
|
|
||
| function getstate() | ||
| state = StateDummy(convert($T, 1), x, similar(x), convert($T, NaN), p) | ||
| end | ||
| # Test HagerZhang I0 | ||
| ls = HagerZhang{$T}() | ||
| state = getstate() | ||
| is = InitialHagerZhang{$T}(α0 = convert($T, NaN)) | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test !isnan(state.alpha) | ||
| @test state.alpha isa $T | ||
| @test ls.mayterminate[] == false | ||
|
|
||
| # Test HagerZhang I12 | ||
| ls = HagerZhang{$T}() | ||
| state = getstate() | ||
| is = InitialHagerZhang{$T}(α0 = convert($T, 1)) | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test !isnan(state.alpha) | ||
| @test state.alpha isa $T | ||
| @test ls.mayterminate[] == true | ||
|
|
||
| # Test Static unscaled | ||
| ls = HagerZhang{$T}() | ||
| state = getstate() | ||
| is = InitialStatic{$T}() | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test state.alpha == is.alpha | ||
| @test !isnan(state.alpha) | ||
| @test state.alpha isa $T | ||
| @test ls.mayterminate[] == false | ||
|
|
||
| # Test Static scaled | ||
| ls = HagerZhang{$T}() | ||
| state = getstate() | ||
| is = InitialStatic{$T}(alpha = convert($T, 0.5), scaled = true) | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test !isnan(state.alpha) | ||
| @test !isnan(state.alpha) | ||
| @test state.alpha isa $T | ||
| @test ls.mayterminate[] == false | ||
|
|
||
| # Test Previous | ||
| ls = HagerZhang{$T}() | ||
| state = getstate() | ||
| alpha = state.alpha | ||
| ls.mayterminate[] = true | ||
| is = InitialPrevious{$T}() | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test state.alpha == alpha | ||
| @test !isnan(state.alpha) | ||
| @test state.alpha isa $T | ||
| @test ls.mayterminate[] == true | ||
|
|
||
| # Test Previous NaN | ||
| ls = HagerZhang{$T}() | ||
| state = getstate() | ||
| state.alpha = convert($T, NaN) | ||
| ls.mayterminate[] = true | ||
| is = InitialPrevious{$T}() | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test state.alpha == is.alpha | ||
| @test state.alpha isa $T | ||
| @test ls.mayterminate[] == true | ||
|
|
||
| # Test Quadratic NaN | ||
| ls = HagerZhang{$T}() | ||
| state = getstate() | ||
| is = InitialQuadratic{$T}() | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test state.alpha == is.α0 | ||
| @test !isnan(state.alpha) | ||
| @test ls.mayterminate[] == false | ||
|
|
||
| # Test Quadratic | ||
| ls = HagerZhang{$T}() | ||
| state = getstate() | ||
| state.f_x_previous = 2*phi_0 | ||
| is = InitialQuadratic{$T}(snap2one=(convert($T, 0.9),convert($T, Inf))) | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test !isnan(state.alpha) | ||
| @test ls.mayterminate[] == false | ||
|
|
||
| # Test Quadratic snap2one | ||
| ls = HagerZhang{$T}() | ||
| state = getstate() | ||
| state.f_x_previous = 2*phi_0 | ||
| is = InitialQuadratic{$T}(snap2one=(convert($T, 0.75),convert($T, Inf))) | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test !isnan(state.alpha) | ||
| @test ls.mayterminate[] == false | ||
|
|
||
| # Test ConstantChange NaN | ||
| ls = HagerZhang{$T}() | ||
| state = getstate() | ||
| is = InitialConstantChange{$T}() | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test state.alpha == is.α0 | ||
| @test !isnan(state.alpha) | ||
| @test ls.mayterminate[] == false | ||
|
|
||
| # Test ConstantChange | ||
| ls = HagerZhang{$T}() | ||
| state = getstate() | ||
| is = InitialConstantChange{$T}() | ||
| is.dϕ_0_previous[] = convert($T, 0.1)*dphi_0 | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test !isnan(state.alpha) | ||
| @test ls.mayterminate[] == false | ||
|
|
||
| # Test ConstantChange snap2one | ||
| ls = HagerZhang{$T}() | ||
| state = getstate() | ||
| is = InitialConstantChange{$T}(snap2one=(convert($T, 0.25),convert($T, 1))) | ||
| is.dϕ_0_previous[] = convert($T, 0.1)*dphi_0 | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test !isnan(state.alpha) | ||
| @test ls.mayterminate[] == false | ||
| end | ||
| doublefloatstypes = [Double64, Double32, Double16] | ||
|
|
||
| @testset "Arbitrary precision - initial step guess: $T" for T in | ||
| [Float32, Float64, BigFloat, doublefloatstypes...] | ||
|
|
||
| f(x) = dot(x, x) | ||
| function g!(out, x) | ||
| out[:] = 2x | ||
| end | ||
|
|
||
| x = convert.(T, [-1, -1]) | ||
| df = NLSolversBase.OnceDifferentiable(f,g!,x) | ||
|
|
||
| phi_0, grtmp = NLSolversBase.value_gradient!(df, x) | ||
| @test !isnan(phi_0) | ||
| @test phi_0 isa T | ||
| @test !any(isnan, grtmp) | ||
| @test grtmp isa Vector{T} | ||
| p = -grtmp | ||
| dphi_0 = dot(p, grtmp) | ||
| @test !isnan(dphi_0) | ||
| @test dphi_0 isa T | ||
|
|
||
| function getstate() | ||
| state = StateDummy(convert(T, 1), x, similar(x), convert(T, NaN), p) | ||
| end | ||
| # Test HagerZhang I0 | ||
| ls = HagerZhang{T}() | ||
| state = getstate() | ||
| is = InitialHagerZhang{T}(α0 = convert(T, NaN)) | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test !isnan(state.alpha) | ||
| @test state.alpha isa T | ||
| @test ls.mayterminate[] == false | ||
|
|
||
| # Test HagerZhang I12 | ||
| ls = HagerZhang{T}() | ||
| state = getstate() | ||
| is = InitialHagerZhang{T}(α0 = convert(T, 1)) | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test !isnan(state.alpha) | ||
| @test state.alpha isa T | ||
| @test ls.mayterminate[] == true | ||
|
|
||
| # Test Static unscaled | ||
| ls = HagerZhang{T}() | ||
| state = getstate() | ||
| is = InitialStatic{T}() | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test state.alpha == is.alpha | ||
| @test !isnan(state.alpha) | ||
| @test state.alpha isa T | ||
| @test ls.mayterminate[] == false | ||
|
|
||
| # Test Static scaled | ||
| ls = HagerZhang{T}() | ||
| state = getstate() | ||
| is = InitialStatic{T}(alpha = convert(T, 0.5), scaled = true) | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test !isnan(state.alpha) | ||
| @test !isnan(state.alpha) | ||
| @test state.alpha isa T | ||
| @test ls.mayterminate[] == false | ||
|
|
||
| # Test Previous | ||
| ls = HagerZhang{T}() | ||
| state = getstate() | ||
| alpha = state.alpha | ||
| ls.mayterminate[] = true | ||
| is = InitialPrevious{T}() | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test state.alpha == alpha | ||
| @test !isnan(state.alpha) | ||
| @test state.alpha isa T | ||
| @test ls.mayterminate[] == true | ||
|
|
||
| # Test Previous NaN | ||
| ls = HagerZhang{T}() | ||
| state = getstate() | ||
| state.alpha = convert(T, NaN) | ||
| ls.mayterminate[] = true | ||
| is = InitialPrevious{T}() | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test state.alpha == is.alpha | ||
| @test state.alpha isa T | ||
| @test ls.mayterminate[] == true | ||
|
|
||
| # Test Quadratic NaN | ||
| ls = HagerZhang{T}() | ||
| state = getstate() | ||
| is = InitialQuadratic{T}() | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test state.alpha == is.α0 | ||
| @test !isnan(state.alpha) | ||
| @test ls.mayterminate[] == false | ||
|
|
||
| # Test Quadratic | ||
| ls = HagerZhang{T}() | ||
| state = getstate() | ||
| state.f_x_previous = 2*phi_0 | ||
| is = InitialQuadratic{T}(snap2one=(convert(T, 0.9),convert(T, Inf))) | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test !isnan(state.alpha) | ||
| @test ls.mayterminate[] == false | ||
|
|
||
| # Test Quadratic snap2one | ||
| ls = HagerZhang{T}() | ||
| state = getstate() | ||
| state.f_x_previous = 2*phi_0 | ||
| is = InitialQuadratic{T}(snap2one=(convert(T, 0.75),convert(T, Inf))) | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test !isnan(state.alpha) | ||
| @test ls.mayterminate[] == false | ||
|
|
||
| # Test ConstantChange NaN | ||
| ls = HagerZhang{T}() | ||
| state = getstate() | ||
| is = InitialConstantChange{T}() | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test state.alpha == is.α0 | ||
| @test !isnan(state.alpha) | ||
| @test ls.mayterminate[] == false | ||
|
|
||
| # Test ConstantChange | ||
| ls = HagerZhang{T}() | ||
| state = getstate() | ||
| is = InitialConstantChange{T}() | ||
| is.dϕ_0_previous[] = convert(T, 0.1)*dphi_0 | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test !isnan(state.alpha) | ||
| @test ls.mayterminate[] == false | ||
|
|
||
| # Test ConstantChange snap2one | ||
| ls = HagerZhang{T}() | ||
| state = getstate() | ||
| is = InitialConstantChange{T}(snap2one=(convert(T, 0.25),convert(T, 1))) | ||
| is.dϕ_0_previous[] = convert(T, 0.1)*dphi_0 | ||
| is(ls, state, phi_0, dphi_0, df) | ||
| @test !isnan(state.alpha) | ||
| @test ls.mayterminate[] == false | ||
| end |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters