You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
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.
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.