60-minute blitz out of date

[jklaise]

I'm working through the 60-minute blitz, but it appears it is out of date:

• derivative appears to be succeeded by gradient
• Even so, taking the second derivative of the polynomial example doesn't work:

using Flux.Tracker: gradient
f(x) = 3x^2 + 2x + 1
df(x) = gradient(f, x)
df(5) # (32.0 (tracked),)
ddf(x) = gradient(df, x)
ddf(5) # expect 6

Last line results in

ERROR: Function output is not scalar
Stacktrace:
 [1] losscheck(::Tuple{Flux.Tracker.TrackedReal{Float64}}) at /home/janis/.julia/packages/Flux/8XpDt/src/tracker/back.jl:171
 [2] gradient_(::Function, ::Int64) at /home/janis/.julia/packages/Flux/8XpDt/src/tracker/back.jl:72
 [3] #gradient#24 at /home/janis/.julia/packages/Flux/8XpDt/src/tracker/back.jl:182 [inlined]
 [4] gradient at /home/janis/.julia/packages/Flux/8XpDt/src/tracker/back.jl:182 [inlined]
 [5] ddf(::Int64) at ./REPL[69]:1
 [6] top-level scope at none:0

[thebhatman]

@jklaise I don't think higher order derivatives can be calculated this way in Flux. IMO, it's best to use Hessians to obtain second order derivatives. You can do something like this:

julia> using Flux

julia> using ReverseDiff

julia> f(x) = @. 3*x^2 + 2*x + 1
f (generic function with 1 method)

julia> ReverseDiff.hessian(f,[5])
1×1 Array{Int64,2}:
 6

[MikeInnes]

You can do this but you'll want to follow the syntax given in the latest version of the docs.

We should update the tutorial, but FWIW you should be able to activate the project and get a compatible version of Flux.

[jklaise]

@MikeInnes thanks for the pointer, I think I got confused between several different versions of the API. I'll see if I can finish the blitz successfully and update the tutorial.

[aminya]

This was fixed by #143 . This can be closed