Original Summary

Upfront caveat: it&#x27;s English only! but I&#x27;m personally still just amazed that 34MB (fine-tuned BAAI&#x2F;bge-small-en-v1.5 ) can fit so much knowledge that it&#x27;s possible to have an iPhone 15 running these classifications entirely client-side on arbitrary english text in 10-30ms.<p>The game is a little trivial but thought it was a cool showcase, you type messages to the guard and they get classified as either friendly, bribe, threats, etc.<p>I&#x27;ve been playing a lot with Jev use-cases since it came out and then learned about fine-tuning my own models.<p>The way I went about it was just using Jev first via a proxy to record my calls and store them as my own training data to make running training easier. I built a whole oss platform around doing that with models (of various sizes) <a href="https:&#x2F;&#x2F;github.com&#x2F;doppsh&#x2F;dopp" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;doppsh&#x2F;dopp</a>.<p>Would love to hear any feedback &#x2F; discuss more on either that whole training process or the game itself.<p>Also I did get a ton of help from Opus doing all this, so some things might be a little rough around some edges but plenty of human time has been put in too. Very welcoming of any feedback&#x2F;criticism

  • 情报分类：硬件与数码
  • 分类依据：内容涉及硬件、数码产品或通信卡
  • 信息来源：Hacker News 新项目
  • 发布时间：2026/10/1 02:55:17