现代信息科技2025，Vol.9Issue(22)：12-16,5.DOI:10.19850/j.cnki.2096-4706.2025.22.003
基于LangChain+Deepseek的智慧燃气GIS领域问答系统研究与实现
Research and Implementation of an Intelligent Q&A System for the Smart Gas GIS Domain Based on LangChain+DeepSeek
石善忠1
作者信息
- 1. 上海城市地理信息系统发展有限公司,上海 200032
- 折叠
摘要
Abstract
To address the common issue of low accuracy in vertical applications of Large Language Models(LLMs),this paper adopts LangChain as the core development framework and integrates DeepSeek's advanced LLM technology.By comparing and implementing cutting-edge model deployment solutions such as Ollama and vLLM,the scheme ultimately adopts vLLM to achieve high-performance inference in production environments and effectively resolves Ollama's performance bottlenecks in concurrent scenarios.Through a modular architecture design,this paper develops a comprehensive technical framework that integrates an intent analysis engine,intelligent memory system,multi-agent collaboration,input/output processing chain,and customizable tool integration.To address the complex business scenarios of smart gas GIS systems,this framework is utilized to create a high-performance intelligent dialogue system with deep industry understanding.关键词
LangChain/DeepSeek/Ollama/对话系统Key words
LangChain/DeepSeek/Ollama/dialogue system分类
计算机与自动化引用本文复制引用
石善忠..基于LangChain+Deepseek的智慧燃气GIS领域问答系统研究与实现[J].现代信息科技,2025,9(22):12-16,5.
