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 #901 from JustVugg/fix/tune-arch-guard
Vincenzo Fornaro committed 3 months ago
47a749d
Merge pull request #896 from JustVugg/feat/web-markdown
Vincenzo Fornaro committed 3 months ago
6ff5765
Merge pull request #895 from JustVugg/ci/vulkan-lavapipe
Vincenzo Fornaro committed 3 months ago
febdba8
Merge pull request #899 from JustVugg/diag/pin-budget
Vincenzo Fornaro committed 3 months ago
549f9f1
deepseek_v4: honor CTX in the generate CLI
Blake committed 3 months ago
71ef303
fix: make placement accounting backend neutral
Mario Gutierrez committed 3 months ago
39e5f0c
experiment: add offline residency simulator
Mario Gutierrez committed 3 months ago
cc18bfb
0e5510b
Merge pull request #587 from RDouglasSharp/metal-attn-fmt4
Vincenzo Fornaro committed 3 months ago
98cab3b
Merge pull request #886 from shujaanazhar/olmoe-serve-mode
Vincenzo Fornaro committed 3 months ago
77619b0
Merge pull request #836 from boblaublaw/fmt_2_m1u
Vincenzo Fornaro committed 3 months ago
a126067
e3bccc4
95fdfda
Merge pull request #880 from JustVugg/test/release-dispatch
Vincenzo Fornaro committed 3 months ago
8d06bd1
207532f
fix(cli): dispatch OLMoE one-shot runs
Mario Gutierrez committed 3 months ago
54dea3a