k3-mix-inference alexdolbun · COMPLETED
Research prototype: multi-GPU MoE expert RAM-offload / movement on vLLM (Kimi-K3 MXFP4). Incomplete.
github.com/alexdolbun/k3-mix-inference · ★ 0 · Forks 1 · Size 513 KB
SUMMARY
Technologies 6
Scored 6
Observed 6
Practices 6
Evidence 6
Skips 0
COVERAGE
Analyzed 102 files · 1 commits · 0 API calls
TECHNOLOGIES & DEPTH
JSON LANGUAGE Depth 70
8 files · PRODUCTION
Shell LANGUAGE Depth 70
7 files · PRODUCTION
Markdown LANGUAGE Depth 70
23 files · PRODUCTION
Python LANGUAGE Depth 70
60 files · PRODUCTION
NumPy LIBRARY Depth 70
4 files · PRODUCTION
TypeScript LANGUAGE Depth 74
26 files · PRODUCTION
PRACTICES
documentation · observedautomated_tests · observedcontinuous_integration · absentcontainerization · absentlinting · absentformatting · absent
ACTIVITY & OWNERSHIP
First commit 2026-07-30
Last commit 2026-07-30
Active months 1
Commits 1