TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

AirLLM enables 8GB MacBook run 70B LLM

79 pointsby SlavikCAover 1 year ago

6 comments

rahimnathwaniover 1 year ago
The acknowledgements section in the README links to notebook, which is where OP sourced the techniques:<p><a href="https:&#x2F;&#x2F;www.kaggle.com&#x2F;code&#x2F;simjeg&#x2F;platypus2-70b-with-wikipedia-rag" rel="nofollow">https:&#x2F;&#x2F;www.kaggle.com&#x2F;code&#x2F;simjeg&#x2F;platypus2-70b-with-wikipe...</a><p>The notebook might be an easier read than the repo, but I haven&#x27;t read either yet.<p>EDIT: It&#x27;s very slow, according to the comments in this thread by people who tried it: <a href="https:&#x2F;&#x2F;github.com&#x2F;oobabooga&#x2F;text-generation-webui&#x2F;issues&#x2F;4754">https:&#x2F;&#x2F;github.com&#x2F;oobabooga&#x2F;text-generation-webui&#x2F;issues&#x2F;47...</a>
jugover 1 year ago
Could 2024 become a crisis for commercial AI?<p>1. We&#x27;re only barely getting started with free MOE models and Mistral has already impressed.<p>2. Cloud AI is a poor fit for corporate use at least in EU due to GDPR, the NIS directive and more. You really dont want to exit the EU in where the data processing takes place.<p>3. There are indications of diminishing returns in LLM performance, where a one year later shot at it from Google, despite massive resources in terms of both experts and data set, still doesn&#x27;t have Gemini Pro clearly surpass GPT 3.5 and Ultra probably not GPT 4. Meanwhile, competition like Mistral is closing in.<p>4. The NY Times lawsuit seems like a harbinger for what is to become a theme for AI companies in 2024. Open collaborations are harder to target as legal entities and there is not nearly as much money to gain if you win.<p>All this points toward a a) convergence of performance that will b) be in the benefit of open models to me.<p>Interesting times anyway especially as we are STILL only getting started.
评论 #38800803 未加载
评论 #38799787 未加载
评论 #38799876 未加载
评论 #38800867 未加载
mdrznover 1 year ago
Whoa, I don&#x27;t understand enough to figure out if this is real and scalable or not, but if this is true it&#x27;s a HUGE step forward. Can&#x27;t wait to try and run a 70b LLM on my 32GB RAM desktop w&#x2F; Windows.
erikawwover 1 year ago
This has mixtral support! Can&#x27;t wait to see the next wave of local MOE models. Perhaps cheap fast and local GPT-4 performance is not too far off.
ceeamover 1 year ago
At what SSD wear rate?
评论 #38801992 未加载
great_psyover 1 year ago
I did not dig too deep in the technicalities of it, but is there anything that would stop openAI from also implementing something like this ?<p>Presumably any advances open source community makes towards running on cheap hardware, will also massively benefit the big guys.