---
id: 20261024-T0-05
title: "MoEZip：按路由压缩KV缓存，降低MoE模型长文本显存占用"
title_en: "MoEZip Compresses KV Cache by Routing to Cut MoE Memory Use"
url: https://ai.daily.yangsir.net/daily/20261024-T0-05
issue_date: 2026-10-24
publish_date: 2026-10-23T04:00:00.000Z
category: research
source_name: "arXiv cs.LG (ML)"
source_url: https://arxiv.org/abs/2610.30206
---

# MoEZip：按路由压缩KV缓存，降低MoE模型长文本显存占用

arXiv 新论文提出 MoEZip，针对稀疏混合专家（MoE）大模型的长上下文推理显存瓶颈。MoE 模型通过每个 token 只激活少量参数来控制算力，但长上下文场景下 KV 缓存随序列长度线性增长，显存成为主要限制。MoEZip 的做法是结合路由信息对 KV 缓存做压缩，利用专家激活模式判断哪些缓存内容更值得保留。实际影响：MoE 模型在长文本推理时可以用更少显存支撑更长上下文，降低部署成本。

## English Version

**MoEZip Compresses KV Cache by Routing to Cut MoE Memory Use**

A new arXiv paper introduces MoEZip to address the memory bottleneck in long-context inference for sparse Mixture-of-Experts (MoE) models. MoE models keep compute in check by activating only a small subset of parameters per token, but KV cache grows linearly with sequence length, making memory the main constraint. MoEZip compresses the KV cache using routing information, leveraging expert activation patterns to decide which cached content is worth keeping. Practical impact: MoE models can support longer contexts with less memory, lowering deployment costs for long-text inference.

---

**来源**：[arXiv cs.LG (ML)](https://arxiv.org/abs/2610.30206)

**详情页**：https://ai.daily.yangsir.net/daily/20261024-T0-05

---

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