基于LoRA方法的大语言模型教学模式在高职院校通信专业实践教学中的应用探究  

Study on the Application of LoRA-based Large Language Model Teaching Mode in Practical Teaching of Communication Majors in Vocational Colleges

作　　者：马颂尧 刘佳骥 杨晨 刘智谦 李文 Ma Songyao;Liu Jiaji;Yang Chen;Liu Zhiqian(Chinese People's Armed Police Force Non commissioned Officer School,Hangzhou 311400)

机构地区：[1]中国人民武装警察部队士官学校,浙江杭州311400 [2]不详

出　　处：《中阿科技论坛(中英文)》2025年第3期132-136,共5页China-Arab States Science and Technology Forum

基　　金：武警士官学校校级理论研究课题“基于半实物仿真模拟的维修课程改革研究”(2023KT018)。

摘　　要：在高职院校通信专业实践教学过程中,存在人装配比受限、理论与实践结合困难等问题。对此,文章结合实操实践教学需求,提出了一种基于LoRA方法(低功耗广域网Low-Rate Wide Area Network)的大语言模型教学模式。该模式以教材和训练手册等资料为原始数据,以国产大语言模型为支撑。通过多维度对比分析多种训练方法,发现LoRA方法在确保正确率的同时,所需的数据资源更少且处理时间更短,因此,可选用LoRA微调方法对大语言模型进行强化训练。在授课过程中,基于LoRA方法的大语言模型教学模式可依托电脑作为实践载体,接入使用LoRA方法不断强化训练大语言模型,实现人机交互的自主学习、案例排障、理论学习、模拟训练、测试组卷等功能。该教学模式有助于提高高职院校通信专业实践教学质量。In the practical teaching of communication majors in vocational colleges,there are problems such as limited personnel-to-equipment ratio and difficulties in combining theory with practice.In response to this,the article proposes a LoRA-based large language model teaching mode which is in line with current teaching demands.This model is based on raw data such as textbooks and training manuals,supported by domestic large language models.Through multidimensional comparisons between various training methods,it was found that the LoRA method requires less data resources and shorter processing time while ensuring accuracy.Hence,the LoRA fine-tuning method can be selected for reinforcement training of large language models.During the teaching process,the LoRA method uses computers to continuously strengthen the training of large language models,thus achieving autonomous learning,case troubleshooting,theoretical learning,simulation training,and test paper generation through human-computer interaction.In summary,this teaching mode helps to boost the quality of practical teaching in communication majors in vocational colleges.

关 键 词：大语言模型 教学设计 模型微调 实践教学 高职院校 LoRA方法 

分 类 号：G434[文化科学—教育学] G642.0[文化科学—教育技术学] TP312[自动化与计算机技术—计算机软件与理论]