Merge pull request #347 from AayushSabharwal/as/forwarddiff

fix: implement ForwarDiff.extract_derivative for AbstractVectorOfArray

Author: ChrisRackauckas

Project.toml

@@ -19,6 +19,7 @@ Tables = "bd369af6-aec1-5ad0-b16a-f7cc5008161c"
 
 [weakdeps]
 FastBroadcast = "7034ab61-46d4-4ed7-9d0f-46aef9175898"
+ForwardDiff = "f6369f11-7733-5829-9624-2563aa707210"
 Measurements = "eff96d63-e80a-5855-80a2-b1b0885c5ab7"
 MonteCarloMeasurements = "0987c9cc-fe09-11e8-30f0-b96dd679fdca"
 Tracker = "9f7883ad-71c0-57eb-9f7f-b5c9e6d3789c"
@@ -27,6 +28,7 @@ ReverseDiff = "37e2e3b7-166d-5795-8a7a-e32c996b4267"
 
 [extensions]
 RecursiveArrayToolsFastBroadcastExt = "FastBroadcast"
+RecursiveArrayToolsForwardDiffExt = "ForwardDiff"
 RecursiveArrayToolsMeasurementsExt = "Measurements"
 RecursiveArrayToolsMonteCarloMeasurementsExt = "MonteCarloMeasurements"
 RecursiveArrayToolsReverseDiffExt = ["ReverseDiff", "Zygote"]

ext/RecursiveArrayToolsForwardDiffExt.jl

@@ -0,0 +1,10 @@
+module RecursiveArrayToolsForwardDiffExt
+
+using RecursiveArrayTools
+using ForwardDiff
+
+function ForwardDiff.extract_derivative(::Type{T}, y::AbstractVectorOfArray) where {T}
+    ForwardDiff.extract_derivative.(T, y)
+end
+
+end

test/adjoints.jl

@@ -62,6 +62,10 @@ function loss8(x)
     return sum(abs2, res)
 end
 
+function loss9(x)
+    return VectorOfArray([collect(3i:3i+3) .* x for i in 1:5])
+end
+
 x = float.(6:10)
 loss(x)
 @test Zygote.gradient(loss, x)[1] == ForwardDiff.gradient(loss, x)
@@ -72,3 +76,4 @@ loss(x)
 @test Zygote.gradient(loss6, x)[1] == ForwardDiff.gradient(loss6, x)
 @test Zygote.gradient(loss7, x)[1] == ForwardDiff.gradient(loss7, x)
 @test Zygote.gradient(loss8, x)[1] == ForwardDiff.gradient(loss8, x)
+@test ForwardDiff.derivative(loss9, 0.0) == VectorOfArray([collect(3i:3i+3) for i in 1:5])