use new batching approach in TensorFlow

src/backend/tensorflow/tensorflow.jl

@@ -2,6 +2,7 @@ module TF
 
 using ..Flux, Flow, TensorFlow, Juno
 import Flux: accuracy
+import TensorFlow: RawTensor
 import Juno: info
 
 export tf
@@ -63,11 +64,17 @@ function tf(model)
   Model(sess, [input], g, gradients(g, input))
 end
 
-function (m::Model)(args...)
+batch(x) = Batch((x,))
+
+RawTensor(data::Batch) = RawTensor(rawbatch(data))
+
+function (m::Model)(args::Batch...)
   @assert length(args) == length(m.inputs)
-  Flux.unbatch(run(m.session, m.graph, Dict(zip(m.inputs, map(batch, args)))))
+  run(m.session, m.graph, Dict(zip(m.inputs, args)))
 end
 
+(m::Model)(args...) = m(map(batch, args)...)
+
 function Flux.back!(m::Model, Δ, args...)
   @assert length(args) == length(m.inputs)
   # TODO: keyword arguments to `gradients`

src/dims/batching.jl

@@ -1,8 +1,4 @@
-export batch, Batch
-
-# TODO: support the Batch type only
-batch(x) = reshape(x, (1,size(x)...))
-batch(xs...) = vcat(map(batch, xs)...)
+export Batch
 
 immutable Batch{T,S} <: AbstractVector{T}
   data::CatMat{T,S}