6 comments

  • brainless 0 minutes ago
    I am sorry I did not understand all of it. But, would this allow running large MoE on a local network with experts spread out over multiple cheaper GPUs (or even CPUs)? This would perhaps be more useful than over the Internet, within offices for example.
  • ptsneves 1 hour ago
    P2P inference sounds really nice and it would bring us back to the 2000s culture if not for the fact inference is already so cheap. Even so it is an insurance policy if the cloud providers or governments have ideas of further censoring and monitoring chats.
  • snovv_crash 36 minutes ago
    Cool idea. How do you handle temperature in the verification?
    • vforno 6 minutes ago
      Ho thanks for the comment. Verification does not depend on temperature. Expert execution is deterministic (pure matmul). LUMABRI_VERIFY=N re-runs N% of the calls on a second replica and requires byte-identical output. Temperature (and sampling) happens only on the chatter, after the experts return their activations. So it can be any value (0, 0.7, 1.2…) without affecting the verification contract.
  • Scaled 39 minutes ago
    Wasn't a p2p ai model how skynet got started?

    J/k, this looks cool :)

  • RIshabh235 2 hours ago
    great work
    • vforno 1 hour ago
      Thanks for the support!
  • robertJk 17 minutes ago
    [dead]