TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Llamafile – The easiest way to run LLMs locally on your Mac

27 点作者 paolop超过 1 年前

5 条评论

wokwokwok超过 1 年前
Why?<p>It&#x27;s unsafe and it takes all the choice and control away from you.<p>You should, instead:<p>1) Build a local copy of llama.cpp (literally clone <a href="https:&#x2F;&#x2F;github.com&#x2F;ggerganov&#x2F;llama.cpp">https:&#x2F;&#x2F;github.com&#x2F;ggerganov&#x2F;llama.cpp</a> and run &#x27;make&#x27;).<p>2) Download the model version you actually want from hugging face (for example, from <a href="https:&#x2F;&#x2F;huggingface.co&#x2F;TheBloke&#x2F;Mistral-7B-Instruct-v0.1-GGUF" rel="nofollow noreferrer">https:&#x2F;&#x2F;huggingface.co&#x2F;TheBloke&#x2F;Mistral-7B-Instruct-v0.1-GGU...</a>, with the clearly indicated required RAM for each variant)<p>3) Run the model yourself.<p>I&#x27;ll say this explicitly: these llamafile things are stupid.<p>You <i>should not</i> download <i>arbitrary user uploaded binary executables</i> and run them on your local laptop.<p>Hugging face may do it&#x27;s best to prevent people from taking advantage of this (heck, they literally invented safetensors), but long story short: we can&#x27;t have nice things because people suck.<p>If you start downloading random executables from the internet and running them, you will regret it.<p>Just spend the extra 5 minutes to build llama.cpp yourself. It&#x27;s very, very easy to do and many guides already exist for doing exactly that.
评论 #38523524 未加载
senthil_rajasek超过 1 年前
See also,<p>Llamafile is the new best way to run a LLM on your own computer (simonwillison.net)<p><a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=38489533">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=38489533</a><p>And<p><a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=38464057">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=38464057</a>
评论 #38523128 未加载
Akashic101超过 1 年前
I&#x27;d like to train one of the provided LLM&#x27;s with my one data, I heard that RAG can be used for that. Does anyone have any pointers on how this could be achieved with llamafiles all locally on my server?
paolop超过 1 年前
What&#x27;s your experience with open source LLMs like LLaVA 1.5 or Mistral 7B?
评论 #38523569 未加载
评论 #38522994 未加载
aldarisbm超过 1 年前
Why does this keep popping up on here?
评论 #38527843 未加载