Train Your Own LLM from Scratch(github.com)
141 pointsby kristianpaul3 hours ago9 comments
  • JoeDaDude3 minutes ago
    Coincidentally, I just started on Build a Large Language Model (From Scratch), a repo/book/course by Sebastian Raschka [0][1][2]. Maybe it is a good problem to have to have to decide which learning resource to use.

    [0] https://github.com/rasbt/LLMs-from-scratch

    [1] https://www.manning.com/books/build-a-large-language-model-f...

    [2] https://magazine.sebastianraschka.com/p/coding-llms-from-the...

  • antirez5 minutes ago
    Context: he is one of the MLX developers, a skilled ML researcher.
  • jvican2 hours ago
    If you're interested in this resource, I highly recommend checking out Stanford's CS336 class. It covers all this curriculum in a lot more depth, introduces you into a lot of theoretical aspects (scaling laws, intuitions) and systems thinking (kernel optimization/profiling). For this, you have to do the assignments, of course... https://cs336.stanford.edu/
  • NSUserDefaultsan hour ago
    Been doing it since the day I was born. The beginnings were hard but I’m getting there.
  • steveharing16 minutes ago
    The documentation is really helpful enough to get started
  • ofsen35 minutes ago
    This looks like exact copy of this video of andrej karpathy ( https://youtu.be/kCc8FmEb1nY ) but in a writing format, am i wrong ?
  • baalimago2 hours ago
    Train your LM from scratch*

    I doubt you have a machine big enough to make it "Large".

    • mips_avataran hour ago
      You can fully train a 1.6b model on a single 3090. That’s a reasonably big model.
    • nucleardogan hour ago
      Hey now! I've got a half terabyte of RAM at my disposal! I mean, it's DDR4 but... it's RAM!

      And it's paired with 48 processor cores! I mean, they don't even support AVX512 but they can do math!

      I could totally train a LLM! Or at least my family could... might need my kid to pick up and carry on the project.

      But in all seriousness... you either missed the point, are being needlessly pedantic, or are... wrong?

      This is about learning concepts, and the rest of this is mostly moot.

      On the pedantic or wrong notes--What is the documented cut-off for a "large" language model? Because GPT-2 was and is described as a "large" language model. It had 1.5B parameters. You can just about get a consumer GPU capable of training that for about $400 these days.

      • Malcolmlisk14 minutes ago
        Then rewrite the title and call it "learn how to do a non usable llm from scratch"
  • hiroakiaizawa2 hours ago
    Nice. What scale does this realistically reach on a single machine?
    • 17 minutes ago
      undefined
    • lynx9740 minutes ago
      Model: 36L/36H/576D, 144.2M params

      runs on a Blackwell 6000 Max-Q, using 86GB VRAM. Training supposedly takes 3h40m

  • iamnotarobotman2 hours ago
    This looks great for a first introduction to training LLMs, and it looks simple enough to try this locally. Great job!