3 papers
cs.LG2026
FlexMoE: One-for-All Nested Intra-Expert Pruning for MoE Language Models
Fan Mo, Yuxuan Han, Geng Zhang +2
Mixture-of-Experts (MoE) language models scale model ability with sparsely activated experts, making this architecture a standard recipe for modern large models. However, sparse ac…
cs.LG2025
HelixPipe: Efficient Distributed Training of Long Sequence Transformers with Attention Parallel Pipeline Parallelism
Geng Zhang, Shenggan Cheng, Xuanlei Zhao +2
As transformer sequence lengths grow, existing pipeline parallelisms incur suboptimal performance due to the quadratic attention computation and the substantial memory overhead. To…
cs.LG2025
MoNE: Replacing Redundant Experts with Lightweight Novices for Structured Pruning of MoE
Geng Zhang, Yuxuan Han, Yuxuan Lou +3
Mixture-of-Experts (MoE) enables efficient scaling of large language models by activating only a subset of experts per input token. However, deploying MoE-based models incurs signi…