Personal project自己的项目
MoEFlow
The question这个项目在研究什么
Expert parallelism, routing & kernels专家并行、路由与内核
Routed MoE execution, CUDA Graph compatibility and the boundary between layer gains and serving performance.路由 MoE 执行、CUDA Graph 兼容性，以及层级收益与端到端服务性能之间的边界。
New to the topic? Start with a small example. Already familiar? Go straight to the articles.第一次接触，先跑一个小例子；已有基础，直接进入文章。
From the project repository来自项目仓库
Paired baselines & validation配对基线与验证方法 ↗
Read the original methods and conditions alongside the results.结合原始方法与条件阅读结果。
Scope & current evidence研究边界与当前证据Read before applying a result应用结论前请留意
Layer-level and full-model evidence are tracked separately. Stable serving speedup remains an open question in the repository reports.层级实验与完整模型证据分别记录。仓库报告中，服务层面的稳定加速仍是开放问题。
Experimental implementation and research notebook实验实现与研究档案 · 2 working outlines篇研究提纲