-
demonstrative example:
```
struct foo
bar
end
function rrule(::typeof(getproperty), F::T, x::Symbol) where T x.bar, foo(randn(10)));
back(y)
```
paging @oxinabox as per request.
-
Firstly, thanks the developers for the great package !
I tried to use the package with ``Zygote.jl`` (for automatic differentiation) but sadly this failed due to the usage of ``=.`` in the defini…
-
#169 highlights that we need rules for a range of functionality which lives in NNlib.jl -- this is not surprising, and has largely already been done (see e.g. the Enzyme extension). Someone needs to s…
-
For example:
```
julia> [42][]
42
julia> using Nabla
julia> [42][]
ERROR: StackOverflowError:
Stacktrace:
[1] getindex(::Vector{Int64}; kwargs::Base.Iterators.Pairs{Union{}, Union{}, …
-
I've been revisiting https://github.com/FluxML/Zygote.jl/pull/1328 as part of a larger PR, and discovered this behaviour while running https://github.com/FluxML/Zygote.jl/blob/108e5a19d8fa7187f6eaece7…
-
Calling `frule_via_ad(cfg, (NoTangent(), one(x)), f, x)` to work out the derivative works for numbers but not in general. So this path:
https://github.com/JuliaDiff/ChainRules.jl/blob/5855c10bdbe69…
-
Same as https://github.com/FluxML/Zygote.jl/pull/728 except in ChainRules form?
-
Add a page to the docs explaining what to do when differentiation fails:
- ForwardDiff: type genericity, define dual overload
- ReverseDiff: type genericity
- Zygote, ChainRules: no mutation, cus…
-
Zygote now uses ChainRules as backend for rules
-
We should add backends for the following AD/FD packages:
- [x] ForwardDiff
- [x] ReverseDiff
- [x] FiniteDifferences
- [ ] all ChainRules-supporting ADs (see #11, #39)
- [ ] FiniteDiff
- [x] Tra…