‹ BackHN Continuity

Thread

Bonsai 2 27B: Near-Lossless Compression in a 9x Smaller Footprint

589 points · 200 comments · JonSchneider

  1. Aurornis · · focus · HN ↗
    These are small enough that you can run them entirely in the browser <a href="https:&#x2F;&#x2F;huggingface.co&#x2F;spaces&#x2F;webml-community&#x2F;ternary-bonsai-2-webgpu-kernels" rel="nofollow">https:&#x2F;&#x2F;huggingface.co&#x2F;spaces&#x2F;webml-community&#x2F;ternary-bonsai...

    Remember to clear the downloaded weights afterward.

    Like the last model, it&#x27;s amazing they work as well as they do. Use it for any longer task and they fall apart spectacularly and in interesting ways.

    1. outofpaper · · focus · HN ↗
      So you have some fun examples?
      1. Aurornis · · focus · HN ↗
        Maybe I oversold the fun-ness of it. The most common failure mode is that it goes into loops and you come back to find it exhausted the output length without getting anywhere.
Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.