Fork me on GitHub
#data-science
<
2018-06-29
>
alan06:06:57

@blueberry I'm a Numpy user, were you thinking about replicating the same benchmarks in the series with it?

aaelony06:06:57

@blueberry, can you explain a bit why random is defn'd within a let?

(let [splittable-random (SplittableRandom.)]
  (defn random ^double [^double _]
    (.nextDouble ^SplittableRandom splittable-random)))

scriptor07:06:26

@aaelony probably to close over the instance of SplittableRandom. The function needs to reuse the instance every time it's called

👍 4
blueberry08:06:29

@justalanm Yes, the same benchmarks, but not necessarily just that. Everything related to matrix computations with Numpy where speed is important and you think that comparison would be interesting.

alan08:06:01

Ok, I don't have much time on my hands right now, but I'll try to do at least the benchmarks examples

gigasquid15:06:14

I just made an #mxnet channel if anyone is interested in joining

✔️ 8
👏 4