2 ms·

Well, if your only experience is reading python opencv stack overflow posts, then of course...

• gok 6y ago
  I mean in the sense that using ML for a problem often requires just trying a dozen different modeling techniques, then a bunch of a hyper-parameter searching, then a bunch of stochastic tuning…
  • SoSoRoCoCo 6y ago
    Oh. I see what you mean. Yeah, I guess by definition backwards propagation is trial-end-error. Huh, I never thought of it that way. Thanks for clarifying, I thought you were being saucy: my apologies for being snarky.
    • gok 6y ago
      It's HN; we're all here for the snark!