diff --git a/benchmark/benchmarks.jl b/benchmark/benchmarks.jl
deleted file mode 100644
index c99baef853..0000000000
--- a/benchmark/benchmarks.jl
+++ /dev/null
@@ -1,54 +0,0 @@
-# We run the benchmarks using AirspeedVelocity.jl
-
-# To run benchmarks locally, first install AirspeedVelocity.jl:
-# julia> using Pkg; Pkg.add("AirspeedVelocity"); Pkg.build("AirspeedVelocity")
-# and make sure .julia/bin is in your PATH.
-
-# Then commit the changes and run:
-# $ benchpkg Functors --rev=mybranch,master --bench-on=mybranch 
-
-
-using BenchmarkTools: BenchmarkTools, BenchmarkGroup, @benchmarkable, @btime, @benchmark, judge
-using Flux
-using Optimisers: AdamW
-using LinearAlgebra: BLAS
-using Statistics, Random
-
-using CUDA
-
-const SUITE = BenchmarkGroup()
-const BENCHMARK_CPU_THREADS = Threads.nthreads()
-BLAS.set_num_threads(BENCHMARK_CPU_THREADS)
-
-function setup_train_mlp()
-    d_in = 128
-    d_out = 128
-    batch_size = 128
-    num_iters = 10
-    device = gpu_device()
-    
-    model = Dense(d_in => d_out) |> device
-    x = randn(Float32, d_in, batch_size) |> device
-    y = randn(Float32, d_out, batch_size) |> device
-    opt = Flux.setup(AdamW(1e-3), model)
-    for iter in 1:num_iters
-        ŷ = model(x)
-        # g = gradient(m -> Flux.mse(m(x), y), model)[1]
-        # Flux.update!(opt, model, g)
-        @info iter
-        # GC.gc(true)
-        CUDA.pool_status()
-    end
-    CUDA.pool_status()
-end
-
-@time setup_train_mlp()
-
-
-# for _ in 1:10
-#     g = gradient(m -> Flux.mse(m(x), y), model)[1]
-#     Flux.update!(opt, model, g)
-# end
-GC.gc(true)
-CUDA.reclaim()
-CUDA.pool_status()