EPLB
Expert Parallelism Load Balancer
...EPLB solves this by duplicating heavily used experts (redundancy) and then placing those duplicates across GPUs to even out computational load. It uses policies like hierarchical load balancing (grouped experts placed at node and then GPU level) and global load balancing depending on configuration. The logic is implemented in eplb.py and supports predicting placements given estimated expert usage weights. EPLB aims to reduce hot-spotting and ensure more uniform usage of compute resources in large MoE deployments.