add an Enzyme docs page

docs/make.jl

@@ -36,6 +36,7 @@ makedocs(
             "Flat vs. Nested" => "reference/destructure.md",
             "Callback Helpers" => "reference/training/callbacks.md",
             "Gradients -- Zygote.jl" => "reference/training/zygote.md",
+            "Gradients -- Enzyme.jl" => "reference/training/enzyme.md",
             "Transfer Data to GPU -- MLDataDevices.jl" => "reference/data/mldatadevices.md",
             "Batching Data -- MLUtils.jl" => "reference/data/mlutils.md",
             "OneHotArrays.jl" => "reference/data/onehot.md",

docs/src/reference/training/enzyme.md

@@ -0,0 +1,9 @@
+
+# [Automatic Differentiation using Enzyme.jl](@id autodiff-enzyme)
+
+Flux now builds in support for Enzyme.jl
+
+```@docs
+gradient(f, args::Union{EnzymeCore.Const, EnzymeCore.Duplicated}...)
+Flux.withgradient(f, args::Union{EnzymeCore.Const, EnzymeCore.Duplicated}...)
+```

docs/src/reference/training/zygote.md

@@ -4,8 +4,10 @@ CollapsedDocStrings = true
 
 # [Automatic Differentiation using Zygote.jl](@id autodiff-zygote)
 
-Flux re-exports the `gradient` from [Zygote](https://github.com/FluxML/Zygote.jl), and uses this function within [`train!`](@ref Flux.train!) to differentiate the model. Zygote has its own [documentation](https://fluxml.ai/Zygote.jl/dev/), in particular listing some [important limitations](https://fluxml.ai/Zygote.jl/dev/limitations/).
+Flux's `gradient` function uses [Zygote](https://github.com/FluxML/Zygote.jl) by default, and also uses this function within [`train!`](@ref Flux.train!) to differentiate the model.
+Zygote has its own [documentation](https://fluxml.ai/Zygote.jl/dev/), in particular listing some [important limitations](https://fluxml.ai/Zygote.jl/dev/limitations/).
 
+Flux also has support for Enzyme.jl, documented [on its own page](@ref autodiff-enzyme).
 
 ## Explicit style