-
Notifications
You must be signed in to change notification settings - Fork 39
Description
Changes to gradients.jl made in fc5a08e removed any way to reduce the number of function calls made when using forward difference for gradients as compared to central difference. The finite_difference_gradient!
functions assume that f(x) will be provided in the GradientCache
for forward differences, but the only constructor for GradientCache
forces fx = nothing
.
Here are the lines returning the only cache the constructor will give you:
FiniteDiff.jl/src/gradients.jl
Lines 52 to 53 in 9945e7d
GradientCache{Nothing,typeof(_c1),typeof(_c2),typeof(_c3),fdtype, | |
returntype,inplace}(nothing,_c1,_c2,_c3) |
I can think of three ways to fix it:
- There could be a constructor that accepts an input of
fx
returns the a cache with fx, although fc5a08e was trying to get rid of extra constructors. - Instead of using the type
Nothing
for the first parameter of the returned GradientCache, useUnion{Nothing,returntype}
so the user can manually update it later. - Or, don't bother using the cache and change the following line
FiniteDiff.jl/src/gradients.jl
Line 138 in 9945e7d
fx, c1, c2, c3 = cache.fx, cache.c1, cache.c2, cache.c3
to_fx, c1, c2, c3 = cache.fx, cache.c1, cache.c2, cache.c3
and then before this for loop:
FiniteDiff.jl/src/gradients.jl
Lines 144 to 145 in 9945e7d
if fdtype == Val(:forward) @inbounds for i ∈ eachindex(x)
add something along the lines offx = _fx == Nothing ? f(x) : _fx
2 makes the most sense to me, but I'm not sure what the performance implications are.