4 comments

  • brandonpelfrey 20 hours ago
    Having written a slightly more involved version of this recently myself I think you did a great job of keeping this compact while still readable. This style of library requires some design for sure.

    Supporting higher order derivatives was also something I considered, but it’s basically never needed in production models from what I’ve seen.

    • iguana2000 14 hours ago
      Thanks! I agree about the style
  • jerkstate 1 day ago
    Karpathy’s micrograd did it first (and better); start here: https://karpathy.ai/zero-to-hero.html
    • alkh 1 day ago
      Imho, we should let people experiment as much as they want. Having more examples is better than less. Still, thanks for the link for the course, this is a top-notch one
    • iguana2000 1 day ago
      Karpathy's material is excellent! This was a project I made for fun, and hopefully provides a different perspective on how this can look
    • richard_chase 1 day ago
      Harsh.
    • whattheheckheck 1 day ago
      Why is it better
      • forgotpwd16 23 hours ago
        Cleaner, more straightforward, more compact code, and considered complete in its scope (i.e. implement backpropagation with a PyTorch-y API and train a neural network with it). MyTorch appears to be an author's self-experiment without concrete vision/plan. This is better for author but worse for outsiders/readers.

        P.S. Course goes far beyond micrograd, to makemore (transfomers), minbpe (tokenization), and nanoGPT (LLM training/loading).

      • tfsh 1 day ago
        Because it's an acclaimed, often cited course by a preeminent AI Researcher (and founding member of OAI) rather than four undocumented python files.
        • gregjw 1 day ago
          it being acclaimed is a poor measure of success, theres always room for improvement, how about some objective comparisons?
        • geremiiah 23 hours ago
          Ironically the reason Karpathy's is better is because he livecoded it and I can be sure it's not some LLM vomit. Unfortunately, we are now indundated with newbies posting their projects/tutorials/guides in the hopes that doing so will catch the eye of a recuiter and land them a high paying AI job. That's not so bad in itself except for the fact that most of these people are completely clueless and posting AI slop.
          • iguana2000 22 hours ago
            Haha, couldn't agree with you more. This, however, isn't AI slop. You can see in the commit history that this is from 3 years ago
        • nurettin 1 day ago
          Objective measures like branch depth, execution speed, memory use and correctness of the results be damned.
          • CamperBob2 1 day ago
            Karpathy's implementation is explicitly for teaching purposes. It's meant to be taken in alongside his videos, which are pretty awesome.
  • khushiyant 22 hours ago
    Better readme would be way to go
    • CamperBob2 17 hours ago
      In iguana2000's defense, the code is highly self-documenting.

      It arguably reads cleaner than Karpathy's in some respects, as he occasionally gets a little ahead of his students with his '1337 Python skillz.

  • jjzkkj 1 day ago
    HmcKk