Symbolics.jacobian throws error after importing ReversePropagation
Opened this issue · 3 comments
jenkspt commented
This works fine
using Symbolics
@variables x y
Symbolics.jacobian([x + x*y, x^2 + y],[x, y])
But adding this throws an error
using ReversePropagation
Symbolics.jacobian([x + x*y, x^2 + y],[x, y])
ERROR: TypeError: in typeassert, expected Num, got a value of type Int64
Stacktrace:
[1] setindex!(A::Matrix{Num}, x::Int64, i1::Int64)
@ Base ./array.jl:839
[2] jacobian(ops::Vector{Num}, vars::Vector{Num}; simplify::Bool)
@ Symbolics ~/.julia/packages/Symbolics/27xgb/src/diff.jl:325
[3] jacobian(ops::Vector{Num}, vars::Vector{Num})
@ Symbolics ~/.julia/packages/Symbolics/27xgb/src/diff.jl:325
[4] top-level scope
Using julia 1.6.0 on Ubuntu 20.04
dpsanders commented
Ouch. I have no idea what's going on there.
jenkspt commented
Can you reproduce on your end?
dpsanders commented
Sorry for the delay. Yes, I can reproduce.
Presumably ReversePropagation is somehow overriding / redefining something that it shouldn't.