DeepSeek与0Ilama协同的图书馆本地知识库构建

打开文本图片集

关键词：DeepSeek；Ollama；AnythingLLM；大语言模型；本地知识库构建

中图分类号：TP391.1；G250.7 文献标识码：A

文章编号：2096-4706（2025）21-0148-07

DeepSeek and Ollama Collaborative Library Local Knowledge Base Construction

LIWeiwei， XIAXinnan （YunnanUniversityLibrary，Kunming 65oo91，China）

Abstract： DeepSeek provides a new opportunity for the localization deployment and application development of Large Language Model with its powerful reasoning and text understanding capability，lowAPIcosts，and low computing power requirement.This paper focuses onthetechnicalpathand implementation methodof usingDepSek to buildlocal knowledge base，and focuseson howtobuildalibrary local knowledge basewith complete knowledge structure，excellent performance andaccurate output through DeepSeekandOllma.The textcomprehension abilityand hardware requirements of the local knowledgebasearetestedaiming toprovidetechnicaland practicalreferencefortheconstructionoflocalknowledgebaseand institutionalknowledgebasebasedonDeepSeek.ThisresearchhelpslibrariestosiftAIfromtheoreticaldiscusiontopractical implementation.

Keywords：DeepSeek; Ollama; AnythingLLM; Large Language Model; local knowledge base construction

0 引言

DeepSeek在2025年初被推出，凭借其开源开放、超低成本和国产自主研发三大优势，同时全量开源训练代码、数据清洗工具和微调框架，并公布其训练的技术细节，在原本如火如荼的AI赛场引发了轩然大波。（剩余10315字）