diff --git a/src/layers/stateless.jl b/src/layers/stateless.jl
index 23fd1651..4c216672 100644
--- a/src/layers/stateless.jl
+++ b/src/layers/stateless.jl
@@ -49,8 +49,3 @@ function normalise(x::AbstractArray; dims=1)
   σ′ = std(x, dims = dims, mean = μ′, corrected=false)
   return (x .- μ′) ./ σ′
 end
-
-function normalise(x::AbstractArray, dims)
-  Base.depwarn("`normalise(x::AbstractArray, dims)` is deprecated, use `normalise(a, dims=dims)` instead.", :normalise)
-  normalise(x, dims = dims)
-end
diff --git a/src/onehot.jl b/src/onehot.jl
index 333922fa..d32bc278 100644
--- a/src/onehot.jl
+++ b/src/onehot.jl
@@ -124,11 +124,6 @@ onecold(y::AbstractMatrix, labels...) =
 onecold(y::OneHotMatrix, labels...) =
   mapreduce(x -> Flux.onecold(x, labels...), |, y.data, dims = 2, init = 0)
 
-function argmax(xs...)
-  Base.depwarn("`argmax(...)` is deprecated, use `onecold(...)` instead.", :argmax)
-  return onecold(xs...)
-end
-
 # TODO probably still want this as a custom adjoint Zygote
 # onecold(x::TrackedVector, l...) = onecold(data(x), l...)
 # onecold(x::TrackedMatrix, l...) = onecold(data(x), l...)
diff --git a/src/optimise/train.jl b/src/optimise/train.jl
index bd965f00..6cc4efcf 100644
--- a/src/optimise/train.jl
+++ b/src/optimise/train.jl
@@ -1,6 +1,5 @@
 using Juno
 import Zygote: Params, gradient
-import Base.depwarn
 
 function update!(opt, x, x̄)
   update!(x, -apply!(opt, x, x̄))
@@ -63,10 +62,6 @@ function train!(loss, ps, data, opt; cb = () -> ())
         loss(d...)
       end
       update!(opt, ps, gs)
-      if cb() == :stop
-        depwarn("Use of `:stop` is deprecated; use `Flux.stop()` instead", :stop)
-        break
-      end
     catch ex
       if ex isa StopException
         break