vllm.distributed.eplb ¶
Expert parallelism load balancer (EPLB).
Modules:
| Name | Description |
|---|---|
async_worker | The async worker that transfers experts in the background. |
eplb_state | Expert parallelism load balancer (EPLB) metrics and states. |
policy | |
rebalance_execute | The actual execution of the rearrangement. |