AI-Hypercomputer / AI-Hypercomputer/JetStream
MOE with JetStream
未关闭
- 主要语言
- Python
- 星标
- 457
- 派生
- 67
- PR 合并指标
- 30 天内没有已合并 PR
描述
Could someone explain or point to a doc that explains how MOE is implemented on Jetstream? Specifically, the all-to-all communications, static vs dynamic, sparse matmuls.
I would like to understand how XLA compiles MOE.
贡献指南
评估
这个 Issue 还没有评估数据。