clean up more rebase mess

src/Flux.jl

@@ -18,7 +18,7 @@ import Optimisers: trainable
 using Random: default_rng
 
 using Zygote, ChainRulesCore
-using Zygote: Params, @adjoint, pullback
+using Zygote: @adjoint, pullback
 using Zygote.ForwardDiff: value
 using EnzymeCore: EnzymeCore
 
@@ -59,10 +59,6 @@ export Chain, Dense, Embedding, EmbeddingBag,
   setup, train!,
   # from Optimsers.jl
   destructure, freeze!, thaw!, adjust!, trainables, update!, trainable,
-  # from Functors.jl
-  functor, @functor, KeyPath, haskeypath, getkeypath,
-  # from Train/Optimisers.jl
-  setup, update!, destructure, freeze!, adjust!, params, trainable, trainables,
   withgradient,
   # init
   glorot_uniform,

src/train.jl

@@ -8,8 +8,6 @@ using ..Flux: Flux
 using ProgressLogging: @progress, @withprogress, @logprogress
 using Zygote: Zygote
 
-# import ..Flux.Optimise: train!, update!, Optimise  # during 0.13, we add methods to the old functions
-
 export setup, train!
 
 using ProgressLogging: @progress, @withprogress, @logprogress
@@ -163,11 +161,6 @@ train!(loss, model::Duplicated, data, opt; cb = nothing) = _enzyme_train!(loss,
 # FluxEnzymeExt defines more specific _enzyme_train!(loss, model::Duplicated, data, opt; cb)
 _enzyme_train!(loss, model, data, opt; cb = nothing) = throw(ArgumentError("The method `train!(loss, Duplicated(model), data, opt_state)` is only available when Enzyme.jl is loaded"))
 
-# # Following src/deprecations.jl
-# function train!(loss, model::Duplicated, data, opt::Optimise.AbstractOptimiser; cb=nothing)
-#   train!(loss, model, data, _old_to_new(opt); cb)
-# end
-
 # This method let you use Optimisers.Descent() without setup, when there is no state
 function train!(loss, model::Duplicated, data, rule::Optimisers.AbstractRule; cb=nothing)
   train!(loss, model, data, _rule_to_state(model, rule); cb)