---
id: 20261104-T0-10
title: "大模型量化新解法：同时处理权重离群值与稀疏性，降低部署内存"
title_en: "Quantizing Large Models with Weight Outliers and Sparsity to Cut Deployment Memory"
url: https://ai.daily.yangsir.net/daily/20261104-T0-10
issue_date: 2026-11-04
publish_date: 2026-11-03T05:00:00.000Z
category: research
source_name: "arXiv cs.LG (ML)"
source_url: https://arxiv.org/abs/2611.00286
---

# 大模型量化新解法：同时处理权重离群值与稀疏性，降低部署内存

研究聚焦大模型训练后量化（PTQ）中的权重离群值与稀疏性问题。PTQ通过把高精度权重转为更低比特来降低部署内存，但会引入量化误差、拖累模型性能。该工作针对权重离群值和稀疏结构提出处理方案，目标是在保持内存效率的同时抑制量化带来的精度损失。对需要在有限显存下部署大模型的团队而言，这类方法直接关系到推理成本。

## English Version

**Quantizing Large Models with Weight Outliers and Sparsity to Cut Deployment Memory**

This work addresses weight outliers and sparsity in post-training quantization (PTQ) for large models. PTQ casts high-precision weights to fewer bits to reduce deployment memory, but introduces quantization error that degrades model performance. The paper proposes handling schemes for weight outliers and sparse structures, aiming to suppress accuracy loss while keeping memory efficiency. For teams deploying large models under tight GPU memory, such methods directly affect inference cost.

---

**来源**：[arXiv cs.LG (ML)](https://arxiv.org/abs/2611.00286)

**详情页**：https://ai.daily.yangsir.net/daily/20261104-T0-10

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*