paper-with-me

홈 › Papers

MxMoE: Mixed-precision Quantization for MoE with Accuracy and Performance Co-Design

2025-05-09 · Haojie Duanmu, Xiuhong Li, Zhihang Yuan, Size Zheng, Jiangfei Duan, Xingcheng Zhang, Dahua Lin

Mixture-of-Experts (MoE) models face deployment challenges due to their large parameter counts and computational demands. We explore quantization for MoE models and highlight two key insights: 1) linear blocks exhibit varying quantization sensitivity, and 2) divergent expert activation frequencies create heterogeneous computational characteristics. Based on these observations, we introduce MxMoE, a mixed-precision optimization framework for MoE models that considers both algorithmic and system perspectives. MxMoE navigates the design space defined by parameter sensitivity, expert activation dynamics, and hardware resources to derive efficient mixed-precision configurations. Additionally, MxMoE automatically generates optimized mixed-precision GroupGEMM kernels, enabling parallel execution of GEMMs with different precisions. Evaluations show that MxMoE outperforms existing methods, achieving 2.4 lower Wikitext-2 perplexity than GPTQ at 2.25-bit and delivering up to 3.4x speedup over full precision, as well as up to 29.4% speedup over uniform quantization at equivalent accuracy with 5-bit weight-activation quantization. Our code is available at https://github.com/cat538/MxMoE.

📄 PDF Abstract BibTeX arXiv:2505.05799

Code (1)

cat538/mxmoe 공식 구현 pytorch

Tasks

Mixture-of-ExpertsQuantizationSensitivity

Methods 이 논문이 사용한 방법론

MoE 설명 없음

Similar Papers 제목 키워드 기반

Q-Strata: Hierarchical Bit Allocation for Mixed-Precision Quantization of Mixture-of-Experts LLMs

2026-08-31 · Deokjae Lee, Sihun Chu, Hyun Oh Song arxiv

Mixed-precision quantization (MPQ) assigns a different bitwidth to each linear layer of a large language model (LLM) to minimize the quantization-induced quality loss under a fixed budget, but Mixture-of-Experts (MoE) mo…

FLIQS: One-Shot Mixed-Precision Floating-Point and Integer Quantization Search

2023-08-07 · Jordan Dotzel, Gang Wu, Andrew Li, Muhammad Umar 외

Quantization has become a mainstream compression technique for reducing model size, computational requirements, and energy consumption for modern deep neural networks (DNNs). With improved numerical support in recent har…

Quantization

CSQ: Growing Mixed-Precision Quantization Scheme with Bi-level Continuous Sparsification

2022-12-06 · Lirui Xiao, Huanrui Yang, Zhen Dong, Kurt Keutzer 외

Mixed-precision quantization has been widely applied on deep neural networks (DNNs) as it leads to significantly better efficiency-accuracy tradeoffs compared to uniform quantization. Meanwhile, determining the exact pre…

Quantization

On-Chip Hardware-Aware Quantization for Mixed Precision Neural Networks

2023-09-05 · Wei Huang, Haotong Qin, Yangdong Liu, Jingzhuo Liang 외

Low-bit quantization emerges as one of the most promising compression approaches for deploying deep neural networks on edge devices. Mixed-precision quantization leverages a mixture of bit-widths to unleash the accuracy …

Quantization

Mixed-Precision Quantization for Deep Vision Models with Integer Quadratic Programming

2023-07-11 · Zihao Deng, Sayeh Sharify, Xin Wang, Michael Orshansky

Quantization is a widely used technique to compress neural networks. Assigning uniform bit-widths across all layers can result in significant accuracy degradation at low precision and inefficiency at high precision. Mixe…

QuantizationSensitivity