KL-MoE: A Hierarchical MoE Pruning Framework Exploiting KL Divergence
Overview
Hierarchical pruning of mixture-of-experts models using KL divergence.
利用 KL 散度进行混合专家模型的分层剪枝。
DAC, 2026. Public paper and code links will be added when available.
Research by Minnan Pei (裴旻楠) and collaborators · Email