Original Summary

Before I get into the details, I&#x27;d just like to say I was able to build this thanks to the work done by the people who work on Pipecat &amp; SmartTurn as this relied on their data.<p>SmartTurn is a fantastic open source model for turn detection. I needed a much smaller version that could fit anywhere and be always on (always actually being every 20ms in this case), so I figured I could do a trade off of making it support one language only (English), not rely on building on top of a Whisper model and possibly squeeze out similar performance.<p>While I do still wish to improve this model significantly, this is my progress in doing so thus far. I also augmented the data with synthetic noise and reverb to try to improve its scores in handling those situations.<p>I&#x27;d love for people to try it, and even better, provide suggestions or audio that can help me improve it.<p>The link takes you to a sample of it running in the browser. There&#x27;s also an SSM version but that was purely for experimental purposes. While SSMs do well on streaming, there isn&#x27;t much to gain from it here, since long memory of audio stream isn&#x27;t that important for turn detection, just the most recent few seconds really, so the GRU version does just fine (or better) and uses much less CPU.<p>More details are in the repo: <a href="https:&#x2F;&#x2F;github.com&#x2F;sshah03&#x2F;CueLight" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;sshah03&#x2F;CueLight</a><p>For those interested in SmartTurn: <a href="https:&#x2F;&#x2F;github.com&#x2F;pipecat-ai&#x2F;smart-turn" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;pipecat-ai&#x2F;smart-turn</a>

  • 情报分类：技术学习与提效
  • 分类依据：内容涉及技术、AI、软件工具或工程实践
  • 信息来源：Hacker News 新项目
  • 发布时间：2026/10/14 23:28:25