@oneiro hadn't seen you on my feed for a while. Glad to see you're still around 😄
@cooler_ranch around & about & sick at home finding new outlets :) how are you?
@oneiro i moved cross country in June with my partner for their job. Enjoying funemployment but using the time to program more. It's weird being a domestic partner.
@cooler_ranch ooh! get to know the neighbourhood, start a small cult, start a small open source project, summon a demon to deal with the out-of-control project ...
@oneiro solid advice. I am learning the neighborhood, am in a cult (organizing), and am working on a FLOSS library.
Haven't needed to summon a demon yet, but I'll add it to the list if things get out of hand.
@cooler_ranch nice! what's the library?
@cooler_ranch ooh. genetic algorithms though...wonder how that compares to the tradeoff of additional epochs on a pretty-good hand-designed network
NN in Erlang Afficher plus
Academic programming Afficher plus
Academic programming Afficher plus
Academic programming Afficher plus
Academic programming Afficher plus
Academic programming Afficher plus
Academic programming Afficher plus
Academic programming Afficher plus
Academic programming Afficher plus
NN in Erlang Afficher plus
@oneiro it will be interesting to see. My first goal is just to give the work a good refactoring. If it's actually applicable, then that's a huge bonus.
The original stuff is written by an academic used to C++ programming who learned enough Erlang to demonstrate the idea. Nothing is idiomatic, there are no tests, the documentation is the book and huge comments that look like:
%%%%%%%%%%:%COMMENT%%%%%%%%%%%
The language/ runtime itself has changed a bit since the publication of the work, too. But biggest of all: the code doesn't really make much use of the fault-tolerant behaviors that are easily used in Erlang and Elixir.