Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Complex numbers handling for GaussAdjoint #996

Open
ArnoStrouwen opened this issue Jan 29, 2024 · 0 comments
Open

Complex numbers handling for GaussAdjoint #996

ArnoStrouwen opened this issue Jan 29, 2024 · 0 comments
Labels

Comments

@ArnoStrouwen
Copy link
Member

using SciMLSensitivity
using OrdinaryDiffEq
using Zygote
dynamics = (x, p, t) -> -x*p[1]

function loss1(params)
    u0 = [1.0+1.0im]
    problem = ODEProblem(dynamics, u0, (0.0, 1.0), params)
    rollout = solve(problem, Tsit5(), u0 = u0, p = params,
        sensealg = InterpolatingAdjoint(autodiff=false, autojacvec=false))
    abs(sum(Array(rollout)[:, end]))
end
function loss2(params)
    u0 = [1.0+1.0im]
    problem = ODEProblem(dynamics, u0, (0.0, 1.0), params)
    rollout = solve(problem, Tsit5(), u0 = u0, p = params,
        sensealg = GaussAdjoint(autodiff=false, autojacvec=false))
    abs(sum(Array(rollout)[:, end]))
end
params = ones(1)
julia> Zygote.gradient(loss1, params)
([-0.5202599109305424],)

julia> Zygote.gradient(loss2, params)
ERROR: MethodError: no method matching (::SciMLBase.ParamJacobianWrapper{false, ODEFunction{…}, Float64, Vector{…}})(::Vector{ComplexF64}, ::Vector{Float64})

Closest candidates are:
  (::SciMLBase.ParamJacobianWrapper{false})(::Any)
   @ SciMLBase ~/.julia/packages/SciMLBase/slQep/src/function_wrappers.jl:87

Stacktrace:
  [1] finite_difference_jacobian!(J::Matrix{…}, f::SciMLBase.ParamJacobianWrapper{…}, x::Vector{…}, cache::FiniteDiff.JacobianCache{…}, f_in::Nothing; relstep::Float64, absstep::Float64, colorvec::UnitRange{…}, sparsity::Nothing, dir::Bool)
    @ FiniteDiff ~/.julia/packages/FiniteDiff/BgLLM/src/jacobians.jl:438
  [2] finite_difference_jacobian!(J::Matrix{ComplexF64}, f::Function, x::Vector{Float64}, cache::FiniteDiff.JacobianCache{Vector{…}, Vector{…}, Vector{…}, Vector{…}, UnitRange{…}, Nothing, Val{…}(), ComplexF64}, f_in::Nothing)
    @ FiniteDiff ~/.julia/packages/FiniteDiff/BgLLM/src/jacobians.jl:341
  [3] finite_difference_jacobian!
    @ FiniteDiff ~/.julia/packages/FiniteDiff/BgLLM/src/jacobians.jl:341 [inlined]
  [4] jacobian!(J::Matrix{…}, f::Function, x::Vector{…}, fx::Vector{…}, alg::GaussAdjoint{…}, jac_config::FiniteDiff.JacobianCache{…})
    @ SciMLSensitivity ~/SciML/SciMLSensitivity.jl/src/derivative_wrappers.jl:157
  [5] vec_pjac!(out::Vector{…}, λ::Vector{…}, y::Vector{…}, t::Float64, S::SciMLSensitivity.GaussIntegrand{…})
    @ SciMLSensitivity ~/SciML/SciMLSensitivity.jl/src/gauss_adjoint.jl:447
  [6] GaussIntegrand
    @ SciMLSensitivity ~/SciML/SciMLSensitivity.jl/src/gauss_adjoint.jl:490 [inlined]
  [7] (::SciMLSensitivity.var"#255#256"{})(out::Vector{…}, u::Vector{…}, t::Float64, integrator::OrdinaryDiffEq.ODEIntegrator{…})
    @ SciMLSensitivity ~/SciML/SciMLSensitivity.jl/src/gauss_adjoint.jl:517
  [8] (::DiffEqCallbacks.SavingIntegrandSumAffect{…})(integrator::OrdinaryDiffEq.ODEIntegrator{…})
    @ DiffEqCallbacks ~/.julia/packages/DiffEqCallbacks/uVI0B/src/integrating_sum.jl:56
  [9] apply_discrete_callback!
    @ ~/.julia/packages/DiffEqBase/eLhx9/src/callbacks.jl:605 [inlined]
 [10] apply_discrete_callback!
    @ ~/.julia/packages/DiffEqBase/eLhx9/src/callbacks.jl:617 [inlined]
 [11] handle_callbacks!(integrator::OrdinaryDiffEq.ODEIntegrator{…})
    @ OrdinaryDiffEq ~/.julia/packages/OrdinaryDiffEq/2nLli/src/integrators/integrator_utils.jl:346
 [12] _loopfooter!(integrator::OrdinaryDiffEq.ODEIntegrator{…})
    @ OrdinaryDiffEq ~/.julia/packages/OrdinaryDiffEq/2nLli/src/integrators/integrator_utils.jl:253
 [13] loopfooter!
    @ ~/.julia/packages/OrdinaryDiffEq/2nLli/src/integrators/integrator_utils.jl:206 [inlined]
 [14] solve!(integrator::OrdinaryDiffEq.ODEIntegrator{…})
    @ OrdinaryDiffEq ~/.julia/packages/OrdinaryDiffEq/2nLli/src/solve.jl:538
 [15] #__solve#746
    @ ~/.julia/packages/OrdinaryDiffEq/2nLli/src/solve.jl:6 [inlined]
 [16] __solve
    @ ~/.julia/packages/OrdinaryDiffEq/2nLli/src/solve.jl:1 [inlined]
 [17] solve_call(_prob::ODEProblem{…}, args::Tsit5{…}; merge_callbacks::Bool, kwargshandle::Nothing, kwargs::@Kwargs{})
    @ DiffEqBase ~/.julia/packages/DiffEqBase/eLhx9/src/solve.jl:609
 [18] solve_call
    @ DiffEqBase ~/.julia/packages/DiffEqBase/eLhx9/src/solve.jl:567 [inlined]
 [19] #solve_up#42
    @ DiffEqBase ~/.julia/packages/DiffEqBase/eLhx9/src/solve.jl:1058 [inlined]
 [20] solve_up
    @ DiffEqBase ~/.julia/packages/DiffEqBase/eLhx9/src/solve.jl:1044 [inlined]
 [21] #solve#40
    @ DiffEqBase ~/.julia/packages/DiffEqBase/eLhx9/src/solve.jl:981 [inlined]
 [22] _adjoint_sensitivities(sol::ODESolution{…}, sensealg::GaussAdjoint{…}, alg::Tsit5{…}; t::Vector{…}, dgdu_discrete::Function, dgdp_discrete::Nothing, dgdu_continuous::Nothing, dgdp_continuous::Nothing, g::Nothing, abstol::Float64, reltol::Float64, checkpoints::Vector{…}, corfunc_analytical::Bool, callback::Nothing, kwargs::@Kwargs{})
    @ SciMLSensitivity ~/SciML/SciMLSensitivity.jl/src/gauss_adjoint.jl:536
 [23] _adjoint_sensitivities
    @ SciMLSensitivity ~/SciML/SciMLSensitivity.jl/src/gauss_adjoint.jl:503 [inlined]
 [24] #adjoint_sensitivities#63
    @ SciMLSensitivity ~/SciML/SciMLSensitivity.jl/src/sensitivity_interface.jl:386 [inlined]
 [25] (::SciMLSensitivity.var"#adjoint_sensitivity_backpass#307"{@Kwargs{}, Tsit5{}, GaussAdjoint{}, Vector{}, Vector{}, SciMLBase.ChainRulesOriginator, Tuple{}, Colon, @NamedTuple{}})(Δ::Matrix{ComplexF64})
    @ SciMLSensitivity ~/SciML/SciMLSensitivity.jl/src/concrete_solve.jl:515
 [26] ZBack
    @ ~/.julia/packages/Zygote/jxHJc/src/compiler/chainrules.jl:211 [inlined]
 [27] (::Zygote.var"#291#292"{Tuple{NTuple{}, Tuple{}}, Zygote.ZBack{SciMLSensitivity.var"#adjoint_sensitivity_backpass#307"{}}})(Δ::Matrix{ComplexF64})
    @ Zygote ~/.julia/packages/Zygote/jxHJc/src/lib/lib.jl:206
 [28] (::Zygote.var"#2169#back#293"{Zygote.var"#291#292"{Tuple{NTuple{}, Tuple{}}, Zygote.ZBack{SciMLSensitivity.var"#adjoint_sensitivity_backpass#307"{}}}})(Δ::Matrix{ComplexF64})
    @ Zygote ~/.julia/packages/ZygoteRules/M4xmc/src/adjoint.jl:72
 [29] #solve#40
    @ ~/.julia/packages/DiffEqBase/eLhx9/src/solve.jl:981 [inlined]
 [30] (::Zygote.Pullback{Tuple{DiffEqBase.var"##solve#40", GaussAdjoint{…}, Vector{…}, Vector{…}, Val{…}, @Kwargs{}, typeof(solve), ODEProblem{…}, Tsit5{…}}, Any})(Δ::Matrix{ComplexF64})
    @ Zygote ~/.julia/packages/Zygote/jxHJc/src/compiler/interface2.jl:0
 [31] #291
    @ Zygote ~/.julia/packages/Zygote/jxHJc/src/lib/lib.jl:206 [inlined]
 [32] (::Zygote.var"#2169#back#293"{Zygote.var"#291#292"{Tuple{NTuple{}, Tuple{}}, Zygote.Pullback{Tuple{}, Any}}})(Δ::Matrix{ComplexF64})
    @ Zygote ~/.julia/packages/ZygoteRules/M4xmc/src/adjoint.jl:72
 [33] solve
    @ ~/.julia/packages/DiffEqBase/eLhx9/src/solve.jl:971 [inlined]
 [34] (::Zygote.Pullback{Tuple{typeof(Core.kwcall), @NamedTuple{…}, typeof(solve), ODEProblem{…}, Tsit5{…}}, Any})(Δ::Matrix{ComplexF64})
    @ Zygote ~/.julia/packages/Zygote/jxHJc/src/compiler/interface2.jl:0
 [35] loss2
    @ ./REPL[7]:4 [inlined]
 [36] (::Zygote.Pullback{Tuple{…}, Tuple{…}})(Δ::Float64)
    @ Zygote ~/.julia/packages/Zygote/jxHJc/src/compiler/interface2.jl:0
 [37] (::Zygote.var"#75#76"{Zygote.Pullback{Tuple{}, Tuple{}}})(Δ::Float64)
    @ Zygote ~/.julia/packages/Zygote/jxHJc/src/compiler/interface.jl:91
 [38] gradient(f::Function, args::Vector{Float64})
    @ Zygote ~/.julia/packages/Zygote/jxHJc/src/compiler/interface.jl:148
 [39] top-level scope
    @ REPL[10]:1
Some type information was truncated. Use `show(err)` to see complete types.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

1 participant