问HN：这里谁在生产环境中使用本地大型语言模型（LLMs）？
我之所以问这个问题，是因为似乎没有人真正这样做。是的，这里那里有一些项目，但最终大家都转向了云端的大型语言模型（LLM）。一切都在云端。人们在偏远的地方为GPU使用付费。但实际上，没有人长期使用本地的LLM。他们说：“哦，这太棒了。本地LLM可以在小设备上运行，甚至可以在你的手机上运行。”
我必须说，对我来说有一个例外，那就是Whisper。我确实经常使用Whisper。但我就是不使用本地的LLM。与云端的GPU相比，它们实在是太差了。
我不知道为什么，因为在我看来，创建一个语音转文本模型比仅仅创建一个生成文本的模型要困难得多。
但似乎他们真的无法消除这些差异，使其在消费者电脑上运行。因此，我还是回到了云端的LLM，尽管隐私问题不在考虑之内。
查看原文
I'm asking because it seems that nobody really does that. Yes, there are some projects here and there, but ultimately everybody just jumps over to cloud LLMs. Everything is cloud. People pay for GPU usage somewhere in the middle of nowhere. But nobody really uses local LLMs long term. They say, "Well, it's so great. Local LLMs work on small devices they even work on your mobile phone."<p>I have to say there's one exception for me and that's Whisper. I actually do use Whisper a lot. But I just don't use local LLMs. They're just really, really bad compared to cloud GPUs.<p>And I don't know why, because for me it seems that having a speech-to-text model is much more challenging to create than just a model that creates text.<p>But it seems that they really cannot remove the differences and have it run on consumer computers. And so I also go back to cloud LLMs, all privacy aside.