NanoGPT Speedrun Frontier

(primeintellect.ai)

28 points | by stared 1 hour ago

1 comments

  • skybrian 5 minutes ago
    Neat!

    The graphs show the "best validated result" for each model. I wonder how much variation there is between runs for a model?