Skip to content

Run frontier MoE models on hardware you already own — pure C, zero deps, experts streamed from disk. Tiny engine, immense model. 🐦

C

Commits on main

Merge pull request #1019 from JustVugg/dev
Vincenzo Fornaro committed 3 months ago
a4d7c77
Merge pull request #1020 from JustVugg/chore/version-1.6.2
Vincenzo Fornaro committed 3 months ago
a99d41c
chore: version 1.6.2
JustVugg committed 3 months ago
1c884f0
7248d53
bbaf161
performance: use coli_v4_route_bf16 in moe_token
Lars Weber committed 3 months ago
042d21f
90a6e1f
12eccb4
6fe6c99
Merge pull request #966 from martinm86867-ops/main
Vincenzo Fornaro committed 3 months ago
c3de36a
76ed063
2494abc
a1648c8
e92f389