About This Resource

A project for running AI models across multiple local devices. It coordinates distributed inference so available compute and memory can be used across a group of machines.

Saved Notes

Original notes saved with this link. They may describe an earlier version.

https://github.com/exo-explore/exo

Forget expensive NVIDIA GPUs, unify your existing devices into one powerful GPU: iPhone, iPad, Android, Mac, Linux, pretty much any device!

GitHub - exo-explore/exo: Run your own AI cluster at home with everyday devices 📱💻 🖥️⌚

Assessment: Trial

Exo has seen good use by the community and is one of the projects that’s trying to do the “I have X computers, let me compute AI across them” scenario.

However, this space is still highly premature (as of December 3, 2024 11:15 AM) and a lot of important bits in Exo does not work yet. Think of it as in “alpha” stage. It’s nowhere near as mature as Ollama or llama.cpp.