NEW REPO

sqliteai/waste — Run the full 2.78-trillion-parameter Kimi K3 model beyond available RAM by streaming activated weights directly from NVMe. A dependency-free, embeddable C inference engine.

Source sqliteaiPublished 3d ago · Aug 1, 2026Posted on Bluesky

What we hold

Stars
482
Language
C

Community signal — surfaced by readers, not stated by the vendor.