Kimi K3 Architecture Overview and Notes

Kimi K3架构概览与笔记

HN 240 分 · 30 条评论 · 作者 ModelForge · 来源 sebastianraschka.com · HN 讨论

【摘要】

Moonshot AI has released the Kimi K3, a massive 2.8T parameter open-weight model that serves as a scaled-up production version of its previous Kimi Linear architecture. This release currently stands as the largest open-weight model available.

⋯ 继续阅读请登录会员 ⋯