[ajtulloch]

A bit petty, but the first example is an unstable softmax implementation:

  defn softmax(t) do
    Nx.exp(t) / Nx.sum(Nx.exp(t))
  end

See https://ogunlao.github.io/2020/04/26/you_dont_really_know_so... etc.

[josevalim]

Thanks for the info and the reference, that was a good read!