Merge branch 'nn-lux' of https://github.com/rbSparky/GraphNeuralNetwo…

…rks.jl into nn-lux

GNNLux/src/layers/conv.jl

@@ -666,25 +666,14 @@ function (l::NNConv)(g, x, edge_weight, ps, st)
     return y, stnew
 end
 
-function LuxCore.initialparameters(rng::AbstractRNG, l::NNConv)
-    weight = l.init_weight(rng, l.out_dims, l.in_dims)
-    if l.use_bias
-        bias = l.init_bias(rng, l.out_dims)
-        return (; weight, bias)
-    else
-        return (; weight)
-    end
-end
-
-LuxCore.parameterlength(l::NNConv) = l.use_bias ? l.in_dims * l.out_dims + l.out_dims : l.in_dims * l.out_dims # nn wont affect this right?
 LuxCore.outputsize(d::NNConv) = (d.out_dims,)
 
-function Base.show(io::IO, l::GINConv)
+function Base.show(io::IO, l::NNConv)
     print(io, "NNConv($(l.nn)")
     print(io, ", $(l.ϵ)")
     l.σ == identity || print(io, ", ", l.σ)
     l.use_bias || print(io, ", use_bias=false")
     l.add_self_loops || print(io, ", add_self_loops=false")
     !l.use_edge_weight || print(io, ", use_edge_weight=true")
     print(io, ")")
-end
+end