For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /get-started.md).

Python function

rebalance_experts_hierarchical

`rebalance_experts_hierarchical()`

max.pipelines.lib.eplb_rebalance.rebalance_experts_hierarchical(weight, num_physical_experts, num_groups, num_nodes, num_gpus)

source

Hierarchical rearrangement: groups -> nodes -> GPUs.

See the reference implementation in utils/benchmarking/speed-of-light/speed_of_light/eplb/rebalance_algo.py for the algorithm walkthrough.

Parameters:

Return type: