removing caching tests

Author: ChrisRackauckas

test/neural_de.jl

@@ -6,7 +6,6 @@ xs = Float32.(hcat([0.; 0.], [1.; 0.], [2.; 0.]))
 tspan = (0.0f0,1.0f0)
 dudt = Flux.Chain(Flux.Dense(2,50,tanh),Flux.Dense(50,2))
 fastdudt = FastChain(FastDense(2,50,tanh),FastDense(50,2))
-fastcdudt = FastChain(FastDense(2,50,tanh,precache=true,numcols=size(xs)[2]),FastDense(50,2,precache=true,numcols=size(xs)[2]))
 
 NeuralODE(dudt,tspan,Tsit5(),save_everystep=false,save_start=false)(x)
 NeuralODE(dudt,tspan,Tsit5(),saveat=0.1)(x)
@@ -68,13 +67,6 @@ gradsnc2 = Zygote.gradient(()->sum(node(xs)),Flux.params(xs,node))
 @test ! iszero(gradsnc2[xs])
 @test ! iszero(gradsnc2[node.p])
 
-nodec = NeuralODE(fastcdudt,tspan,Tsit5(),save_everystep=false,save_start=false,p=pd)
-gradsc = Zygote.gradient(()->sum(nodec(x)),Flux.params(x,nodec)) #with cache
-@test ! iszero(gradsc[x])
-@test ! iszero(gradsc[nodec.p])
-@test gradsnc[x] ≈ gradsc[x] rtol=1e-6
-@test gradsnc[node.p] ≈ gradsc[nodec.p] rtol=1e-6
-
 gradsc2 = Zygote.gradient(()->sum(nodec(xs)),Flux.params(xs,nodec))
 @test ! iszero(gradsc2[xs])
 @test ! iszero(gradsc2[nodec.p])
@@ -87,11 +79,6 @@ gradsnc = Zygote.gradient(()->sum(node(x)),Flux.params(x,node))
 @test ! iszero(gradsnc[x])
 @test ! iszero(gradsnc[node.p])
 
-nodec = NeuralODE(fastcdudt, tspan, Tsit5(), abstol=1e-12, reltol=1e-12, save_everystep=false, save_start=false,p=pd)
-gradsc = Zygote.gradient(()->sum(nodec(x)),Flux.params(x,nodec))
-@test gradsnc[x] ≈ gradsc[x] rtol=1e-3
-@test gradsnc[node.p] ≈ gradsc[nodec.p] rtol=1e-3
-
 node = NeuralODE(fastdudt,tspan,Tsit5(),save_everystep=false,save_start=false,sensealg=TrackerAdjoint())
 grads = Zygote.gradient(()->sum(node(x)),Flux.params(x,node))
 @test ! iszero(grads[x])
@@ -104,11 +91,6 @@ grads = Zygote.gradient(()->sum(node(xs)),Flux.params(xs,node))
 goodgrad = grads[node.p]
 p = node.p
 
-node = NeuralODE(fastcdudt,tspan,Tsit5(),save_everystep=false,save_start=false,sensealg=TrackerAdjoint())
-grads = Zygote.gradient(()->sum(node(x)),Flux.params(x,node))
-@test ! iszero(grads[x])
-@test ! iszero(grads[node.p])
-
 grads = Zygote.gradient(()->sum(node(xs)),Flux.params(xs,node))
 @test ! iszero(grads[xs])
 @test ! iszero(grads[node.p])
@@ -127,23 +109,6 @@ grads = Zygote.gradient(()->sum(node(xs)),Flux.params(xs,node))
 goodgrad2 = grads[node.p]
 @test goodgrad ≈ goodgrad2 # Make sure adjoint overloads are correct
 
-node = NeuralODE(fastcdudt,tspan,Tsit5(),save_everystep=false,save_start=false, sensealg=BacksolveAdjoint(),p=pc)
-grads = Zygote.gradient(()->sum(node(x)),Flux.params(x,node))
-@test ! iszero(grads[x])
-@test ! iszero(grads[node.p])
-
-grads = Zygote.gradient(()->sum(node(xs)),Flux.params(xs,node))
-@test !iszero(grads[xs])
-@test ! iszero(grads[node.p])
-goodgrad2 = grads[node.p]
-@test goodgradc ≈ goodgrad2 rtol=1e-6
-
-grads = Zygote.gradient(()->sum(node(xs)),Flux.params(xs,node))
-goodgrad2 = grads[node.p]
-@test goodgrad ≈ goodgrad2 rtol = 1e-6
-
-@test_throws ErrorException grads = Zygote.gradient(()->sum(node(xs)),Flux.params(xs,node))
-
 @info "Test some adjoints"
 
 # Adjoint
@@ -184,15 +149,6 @@ goodgrad2 = grads[node.p]
     @test_broken ! iszero(grads[xs])
     @test_broken ! iszero(grads[node.p])
 
-    node = NeuralODE(fastcdudt,tspan,Tsit5(),save_everystep=false,save_start=false)
-    grads = Zygote.gradient(()->sum(node(x)),Flux.params(x,node))
-    @test ! iszero(grads[x])
-    @test ! iszero(grads[node.p])
-
-    @test_broken grads = Zygote.gradient(()->sum(node(xs)),Flux.params(xs,node)) isa Tuple
-    @test_broken ! iszero(grads[xs])
-    @test_broken ! iszero(grads[node.p])
-
     node = NeuralODE(fastdudt,tspan,Tsit5(),saveat=0.0:0.1:1.0)
     grads = Zygote.gradient(()->sum(node(x)),Flux.params(x,node))
     @test ! iszero(grads[x])
@@ -202,15 +158,6 @@ goodgrad2 = grads[node.p]
     @test_broken ! iszero(grads[xs])
     @test_broken ! iszero(grads[node.p])
 
-    node = NeuralODE(fastcdudt,tspan,Tsit5(),saveat=0.0:0.1:1.0)
-    grads = Zygote.gradient(()->sum(node(x)),Flux.params(x,node))
-    @test ! iszero(grads[x])
-    @test ! iszero(grads[node.p])
-
-    @test_broken grads = Zygote.gradient(()->sum(node(xs)),Flux.params(xs,node)) isa Tuple
-    @test_broken ! iszero(grads[xs])
-    @test_broken ! iszero(grads[node.p])
-
     node = NeuralODE(fastdudt,tspan,Tsit5(),saveat=0.1)
     grads = Zygote.gradient(()->sum(node(x)),Flux.params(x,node))
     @test ! iszero(grads[x])
@@ -219,15 +166,6 @@ goodgrad2 = grads[node.p]
     @test_broken grads = Zygote.gradient(()->sum(node(xs)),Flux.params(xs,node)) isa Tuple
     @test_broken ! iszero(grads[xs])
     @test_broken ! iszero(grads[node.p])
-
-    node = NeuralODE(fastcdudt,tspan,Tsit5(),saveat=0.1)
-    grads = Zygote.gradient(()->sum(node(x)),Flux.params(x,node))
-    @test ! iszero(grads[x])
-    @test ! iszero(grads[node.p])
-
-    @test_broken grads = Zygote.gradient(()->sum(node(xs)),Flux.params(xs,node)) isa Tuple
-    @test_broken ! iszero(grads[xs])
-    @test_broken ! iszero(grads[node.p])
 end
 
 @info "Test Tracker"
@@ -270,15 +208,6 @@ end
     @test_broken ! iszero(grads[xs])
     @test_broken ! iszero(grads[node.p])
 
-    node = NeuralODE(fastcdudt,tspan,Tsit5(),save_everystep=false,save_start=false,sensealg=TrackerAdjoint())
-    grads = Zygote.gradient(()->sum(node(x)),Flux.params(x,node))
-    @test ! iszero(grads[x])
-    @test ! iszero(grads[node.p])
-
-    @test_broken grads = Zygote.gradient(()->sum(node(xs)),Flux.params(xs,node)) isa Tuple
-    @test_broken ! iszero(grads[xs])
-    @test_broken ! iszero(grads[node.p])
-
     node = NeuralODE(fastdudt,tspan,Tsit5(),saveat=0.0:0.1:1.0,sensealg=TrackerAdjoint())
     grads = Zygote.gradient(()->sum(node(x)),Flux.params(x,node))
     @test ! iszero(grads[x])
@@ -288,15 +217,6 @@ end
     @test_broken ! iszero(grads[xs])
     @test_broken ! iszero(grads[node.p])
 
-    node = NeuralODE(fastcdudt,tspan,Tsit5(),saveat=0.0:0.1:1.0,sensealg=TrackerAdjoint())
-    grads = Zygote.gradient(()->sum(node(x)),Flux.params(x,node))
-    @test ! iszero(grads[x])
-    @test ! iszero(grads[node.p])
-
-    @test_broken grads = Zygote.gradient(()->sum(node(xs)),Flux.params(xs,node)) isa Tuple
-    @test_broken ! iszero(grads[xs])
-    @test_broken ! iszero(grads[node.p])
-
     node = NeuralODE(fastdudt,tspan,Tsit5(),saveat=0.1,sensealg=TrackerAdjoint())
     grads = Zygote.gradient(()->sum(node(x)),Flux.params(x,node))
     @test ! iszero(grads[x])
@@ -305,22 +225,12 @@ end
     @test_broken grads = Zygote.gradient(()->sum(node(xs)),Flux.params(xs,node)) isa Tuple
     @test_broken ! iszero(grads[xs])
     @test_broken ! iszero(grads[node.p])
-
-    node = NeuralODE(fastcdudt,tspan,Tsit5(),saveat=0.1,sensealg=TrackerAdjoint())
-    grads = Zygote.gradient(()->sum(node(x)),Flux.params(x,node))
-    @test ! iszero(grads[x])
-    @test ! iszero(grads[node.p])
-
-    @test_broken grads = Zygote.gradient(()->sum(node(xs)),Flux.params(xs,node)) isa Tuple
-    @test_broken ! iszero(grads[xs])
-    @test_broken ! iszero(grads[node.p])
 end
 
 @info "Test non-ODEs"
 
 dudt2 = Flux.Chain(Flux.Dense(2,50,tanh),Flux.Dense(50,2))
 fastdudt2 = FastChain(FastDense(2,50,tanh),FastDense(50,2))
-fastcdudt2 = FastChain(FastDense(2,50,tanh,numcols=size(xs)[2],precache=true),FastDense(50,2,numcols=size(xs)[2],precache=true))
 NeuralDSDE(dudt,dudt2,(0.0f0,.1f0),SOSRI(),saveat=0.1)(x)
 sode = NeuralDSDE(dudt,dudt2,(0.0f0,.1f0),SOSRI(),saveat=0.0:0.01:0.1)
 
@@ -348,15 +258,6 @@ gradsnc2 = Zygote.gradient(()->sum(sode(xs)),Flux.params(xs,sode))
 @test ! iszero(gradsnc2[sode.p])
 @test ! iszero(gradsnc2[sode.p][end])
 
-sodec = NeuralDSDE(fastcdudt,fastcdudt2,(0.0f0,.1f0),SOSRI(),saveat=0.0:0.01:0.1,p=pd)
-Random.seed!(1234)
-gradsc = Zygote.gradient(()->sum(sodec(x)),Flux.params(x,sodec))
-@test ! iszero(gradsc[x])
-@test ! iszero(gradsc[sodec.p])
-@test ! iszero(gradsc[sodec.p][end])
-@test gradsnc[x] ≈ gradsc[x] rtol=1e-6
-@test gradsnc[sode.p] ≈ gradsc[sodec.p] rtol=1e-6
-
 gradsc2 = Zygote.gradient(()->sum(sodec(xs)),Flux.params(xs,sodec))
 @test_broken gradsc2 isa Tuple
 @test ! iszero(gradsc2[xs])
@@ -365,7 +266,6 @@ gradsc2 = Zygote.gradient(()->sum(sodec(xs)),Flux.params(xs,sodec))
 
 dudt22 = Flux.Chain(Flux.Dense(2,50,tanh),Flux.Dense(50,4),x->reshape(x,2,2))
 fastdudt22 = FastChain(FastDense(2,50,tanh),FastDense(50,4),(x,p)->reshape(x,2,2))
-fastcdudt22 = FastChain(FastDense(2,50,tanh,numcols=size(xs)[2],precache=true),FastDense(50,4,numcols=size(xs)[2],precache=true),(x,p)->reshape(x,2,2))
 NeuralSDE(dudt,dudt22,(0.0f0,.1f0),2,LambaEM(),saveat=0.01)(x)
 
 sode = NeuralSDE(dudt,dudt22,(0.0f0,0.1f0),2,LambaEM(),saveat=0.0:0.01:0.1)
@@ -393,15 +293,6 @@ gradsnc = Zygote.gradient(()->sum(sode(x)),Flux.params(x,sode))
 @test ! iszero(gradsnc[sode.p])
 @test ! iszero(gradsnc[sode.p][end])
 
-sodec = NeuralSDE(fastcdudt,fastcdudt22,(0.0f0,0.1f0),2,LambaEM(),saveat=0.0:0.01:0.1,p=pd)
-Random.seed!(1234)
-gradsc = Zygote.gradient(()->sum(sodec(x)),Flux.params(x,sodec))
-@test ! iszero(gradsc[x])
-@test ! iszero(gradsc[sodec.p])
-@test ! iszero(gradsc[sodec.p][end])
-@test gradsnc[x] ≈ gradsc[x] rtol=1e-6
-@test gradsnc[sode.p] ≈ gradsc[sodec.p] rtol=1e-6
-
 @test_broken gradsc = Zygote.gradient(()->sum(sodec(xs)),Flux.params(xs,sodec))
 @test_broken ! iszero(gradsc[xs])
 @test ! iszero(gradsc[sodec.p])