Unverified50% confidenceFactExact time

Llama 3 70B量化版（Q4精度）仍需约35GB内存

1

Sources

50%

Confidence

Long-term

Relevance

7/23/2026

First Seen

Sources

Hermes Agent接入个人微信实战：全流程部署指南

Related Claims

Unverified以约160GB的模型（如Qwen3.5-397B的UD-Q3_K_XL量化版本）为例，运行透镜需额外约20GB内存80% similar Unverified3B～4B级别模型如Qwen2.5 3B或Llama 3.2 3B在Q4量化下仅需2～2.5GB显存78% similar VerifiedQwen2.5:7B模型Q4量化后文件大小约为4GB78% similar Unverified70B模型在INT4精度下可压缩至约35GB76% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/594381

API

curl https://kongchang.com/api/v1/knowledge/claims/594381

MCP

get_claim(id=594381)