diff --git a/src/Flux.jl b/src/Flux.jl
index 251a14a5f5..5cd17ad0c4 100644
--- a/src/Flux.jl
+++ b/src/Flux.jl
@@ -35,6 +35,7 @@ include("layers/basic.jl")
 include("layers/conv.jl")
 include("layers/recurrent.jl")
 include("layers/normalise.jl")
+include("layers/upsample.jl")
 
 include("data/Data.jl")
 
diff --git a/src/layers/upsample.jl b/src/layers/upsample.jl
new file mode 100644
index 0000000000..d7048fa8cd
--- /dev/null
+++ b/src/layers/upsample.jl
@@ -0,0 +1,61 @@
+"""
+    BilinearUpsample(factor::Tuple{Integer,Integer})
+
+Create an upsampling layer that uses bilinear interpolation.
+
+The width and height dimensions grow by the `factor` tuple.
+
+# Examples
+```jldoctest; setup = :(using Flux: BilinearUpsample; using Random; Random.seed!(0))
+julia> b = Flux.BilinearUpsample((2, 2))
+BilinearUpsample(2, 2)
+
+julia> b(rand(2, 2, 1, 1))
+4×4×1×1 Array{Float64,4}:
+[:, :, 1, 1] =
+ 0.823648  0.658877  0.329336  0.164566
+ 0.845325  0.675933  0.337149  0.167757
+ 0.888679  0.710044  0.352773  0.174138
+ 0.910357  0.7271    0.360586  0.177329```
+"""
+struct BilinearUpsample{T<:Integer}
+    factor::Tuple{T,T}
+end
+
+function (b::BilinearUpsample)(x::AbstractArray)
+    W, H, C, N = size(x)
+
+    newW = W * b.factor[1]
+    newH = H * b.factor[2]
+
+    out = similar(x, (newW, newH, C, N))
+
+    @inbounds for w = 1:newW, h = 1:newH
+        w₀ = (w - 0.5) / b.factor[1] + 0.5
+        h₀ = (h - 0.5) / b.factor[2] + 0.5
+
+        w1 = floor(Int, w₀)
+        w2 = w1 + 1
+        h1 = floor(Int, h₀)
+        h2 = h1 + 1
+
+        i1 = clamp(w1, 1, W)
+        i2 = clamp(w2, 1, W)
+        j1 = clamp(h1, 1, H)
+        j2 = clamp(h2, 1, H)
+
+        @views out[w, h, :, :] =
+            (
+                x[i1, j1, :, :] * (w2 - w₀) * (h2 - h₀) +
+                x[i1, j2, :, :] * (w2 - w₀) * (h₀ - h1) +
+                x[i2, j1, :, :] * (w₀ - w1) * (h2 - h₀) +
+                x[i2, j2, :, :] * (w₀ - w1) * (h₀ - h1)
+            ) / ((w2 - w1) * (h2 - h1))
+    end
+
+    out
+end
+
+function Base.show(io::IO, b::BilinearUpsample)
+    print(io, "BilinearUpsample(", b.factor[1], ", ", b.factor[2], ")")
+end
diff --git a/test/layers/upsample.jl b/test/layers/upsample.jl
new file mode 100644
index 0000000000..d3df3505b0
--- /dev/null
+++ b/test/layers/upsample.jl
@@ -0,0 +1,11 @@
+using Flux: BilinearUpsample
+using Test
+
+@testset "BilinearUpsample" begin
+  @test size(BilinearUpsample((2, 2))(rand(2, 2, 1, 1))) == (4, 4, 1, 1)
+  @test size(BilinearUpsample((3, 3))(rand(2, 2, 1, 1))) == (6, 6, 1, 1)
+  @test size(BilinearUpsample((2, 2))(rand(2, 2, 10, 10))) == (4, 4, 10, 10)
+  @test size(BilinearUpsample((3, 3))(rand(2, 2, 10, 10))) == (6, 6, 10, 10)
+
+  @test_throws BoundsError BilinearUpsample((2, 2))(rand(2, 2))
+end
diff --git a/test/runtests.jl b/test/runtests.jl
index c2ea0715cf..637460de01 100644
--- a/test/runtests.jl
+++ b/test/runtests.jl
@@ -1,6 +1,6 @@
-using Flux 
+using Flux
 using Flux.Data
-using Test 
+using Test
 using Random, Statistics, LinearAlgebra
 using Documenter
 using IterTools: ncycle
@@ -30,6 +30,7 @@ Random.seed!(0)
     include("layers/normalisation.jl")
     include("layers/stateless.jl")
     include("layers/conv.jl")
+    include("layers/upsample.jl")
   end
 
   @testset "CUDA" begin