🚀 You can now achieve GPT-3 level performance on your Mac at 12 tokens/sec using compressed LLaMa 7B and optimized inference with just 4GB of RAM. Join our Discord for more updates: discord.gg/dnufPxmW. #GPT3 #ChatGPT #AGI #LLaMa

Mar 10, 2023 · 3:06 AM UTC

🤝 We're excited to share our progress on making AI inference feasible on personal devices and technology accessible to all. We're thrilled to announce that we will soon be open sourcing the codebase (without any forms 😛). #OpenSource #AGI #DecentralizedAI #AIForAll
🔜 Experience lightning-fast AI on your personal devices! Join our Discord community to get early access and provide valuable feedback on AI APIs that will reside and run on your personal devices. #AGI #LLMs #ChatGPT #AIFuture
🚀 Personalized AI models that respond to email and messages in your own unique style and have knowledge of your past exchanges - all running privately on your device - are on the horizon! Stay tuned for updates. #PersonalAI #Privacy #AGI
🤣 These AI models may also surprise you with a few laughs! Trained on internet data, they've picked up some weird and wacky facts along the way. Get ready for a one-of-a-kind AI experience on your personal devices! #PersonalizedAI #FunFacts #AGI
Replying to @NolanoOrg
On a fucking mac… eliezer was right… its fucking over isnt it
Saved! Here's the compiled thread: mem.ai/p/pnpOOlH8cOHL5ZQbzl9… 🪄 AI-generated summary: "It is now possible to achieve GPT-3 level performance on Macs with just 4GB of RAM. The codebase will soon be open sourced, and users can join the Discord community to get...
Replying to @NolanoOrg
Uhhhhh RAM? Not VRAM? If so this is HUGE.
Same thing on Apple Silicon.
Saved! Here's the compiled thread: mem.ai/p/HapPrj2RApbCwX13pn5… 🪄 AI-generated summary: "It is now possible to achieve GPT-3 level performance on Macs with just 4GB of RAM. The codebase will soon be open sourced, and users can join the Discord community to get...