[{"data":1,"prerenderedAt":762},["ShallowReactive",2],{"content-query-BsMFhavkOw":3},{"_path":4,"_dir":5,"_draft":6,"_partial":6,"_locale":7,"title":8,"description":9,"date":10,"cover":11,"type":12,"body":13,"_type":756,"_id":757,"_source":758,"_file":759,"_stem":760,"_extension":761},"/news/zh/3271","zh",false,"","昇思MindSpore助力电子科技大学人工智能教学，培养AI原生人才","项目实施背景","2024-08-22","https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2024/08/23/6cb6956658d6401784e1e9dc380ebbc8.png","news",{"type":14,"children":15,"toc":753},"root",[16,24,33,38,43,48,60,68,73,81,86,91,99,107,112,117,124,129,134,139,144,149,156,161,169,177,182,187,194,199,204,209,217,222,227,234,239,244,249,254,259,264,269,274,279,284,289,294,299,307,312,317,322,327,332,337,342,347,352,360,365,370,375,380,385,390,395,400,405,410,415,423,428,435,440,448,456,464,469,476,481,489,494,524,532,537,544,552,557,562,567,572,577,582,587,595,603,611,616,623,628,635,640,647,652,657,664,669,674,697,705,710,715,723,728,733,738,743,748],{"type":17,"tag":18,"props":19,"children":21},"element","h1",{"id":20},"昇思mindspore助力电子科技大学人工智能教学培养ai原生人才",[22],{"type":23,"value":8},"text",{"type":17,"tag":25,"props":26,"children":27},"p",{},[28],{"type":17,"tag":29,"props":30,"children":31},"strong",{},[32],{"type":23,"value":9},{"type":17,"tag":25,"props":34,"children":35},{},[36],{"type":23,"value":37},"发展新质生产力对人工智能人才培养提出了迫切的需求。国内众多高校均已设立人工智能专业，但人工智能创新人才培养模式还面临诸多问题，其中实验教学的问题尤为突出。",{"type":17,"tag":25,"props":39,"children":40},{},[41],{"type":23,"value":42},"实验教学是培养人工智能实践和创新能力的关键环节。当前，人工智能实验内容仍以传统的机器学习算法为主，而作为人工智能技术主要实现途径的深度学习课程的实践却难以开展，普遍存在实验设备和资源不足、课程理论性过强、实践性不足等问题。此外，产学研合作不够深入，课程难以跟踪迅速发展的人工智能产业前沿，导致人才培养与产业脱节。",{"type":17,"tag":25,"props":44,"children":45},{},[46],{"type":23,"value":47},"本项目针对以上问题，建设了面向行业发展需求的AI实验平台与内容，大幅降低了实验平台成本，有效解决了人工智能实践教学的痛点问题，对于人工智能创新人才的培养具有重要意义。",{"type":17,"tag":25,"props":49,"children":50},{},[51,53,58],{"type":23,"value":52},"本项目相关成果入选",{"type":17,"tag":29,"props":54,"children":55},{},[56],{"type":23,"value":57},"2023年第十届教育部产学合作协同育人对接会典型项目成果展，并获教育部产学合作协同育人项目华为优秀成果奖、第五届中国计算机教育大会计算机类教学资源建设优秀教学案例一等奖",{"type":23,"value":59},"等。",{"type":17,"tag":25,"props":61,"children":62},{},[63],{"type":17,"tag":64,"props":65,"children":67},"img",{"alt":7,"src":66},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2024/08/23/6c9117d59d3e4b40b1722535023549ee.png",[],{"type":17,"tag":25,"props":69,"children":70},{},[71],{"type":23,"value":72},"图1. 教育部高等教育司司长周天华（右二）和产学合作协同育人项目专家组组长徐晓飞（右一）在参观本成果展示",{"type":17,"tag":25,"props":74,"children":75},{},[76],{"type":17,"tag":29,"props":77,"children":78},{},[79],{"type":23,"value":80},"改革思路及举措",{"type":17,"tag":25,"props":82,"children":83},{},[84],{"type":23,"value":85},"深度学习实践教学难以开展的根本原因，在于实验硬件平台昂贵、软件平台混乱，难以开展人工智能创新实验。本案例的改革思路是，结合科研项目经验和华为昇腾AI全栈全场景产业技术资源，建设跟踪产业发展前沿、成本低、易于部署和推广的AI实验平台和逐级推进的全流程深度学习实验项目内容，有效解决深度学习的实践教学难题。",{"type":17,"tag":25,"props":87,"children":88},{},[89],{"type":23,"value":90},"本案例的建设目标是，支持人工智能课程的实验教学，通过理论与实践相结合，帮助学生深入了解人工智能的基本原理和应用场景，掌握AI行业前沿技术的基本使用，提高AI创新能力和解决实际问题的能力，培养具有创新精神、掌握前沿技术、具备跨学科融合能力、能够解决实际问题的人工智能创新人才。主要举措如下：",{"type":17,"tag":25,"props":92,"children":93},{},[94],{"type":17,"tag":29,"props":95,"children":96},{},[97],{"type":23,"value":98},"01",{"type":17,"tag":25,"props":100,"children":101},{},[102],{"type":17,"tag":29,"props":103,"children":104},{},[105],{"type":23,"value":106},"实验平台建设",{"type":17,"tag":25,"props":108,"children":109},{},[110],{"type":23,"value":111},"基于华为Atlas 200 DK开发者套件建设成本低、算力强的深度学习实验平台，具体内容可分为硬件平台和软件平台两方面。",{"type":17,"tag":25,"props":113,"children":114},{},[115],{"type":23,"value":116},"（1）硬件平台：Atlas 200 DK开发者套件和集群服务器。",{"type":17,"tag":25,"props":118,"children":119},{},[120],{"type":17,"tag":64,"props":121,"children":123},{"alt":7,"src":122},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2024/08/23/4dd9e81659a74b17836778305231d22f.png",[],{"type":17,"tag":25,"props":125,"children":126},{},[127],{"type":23,"value":128},"图2. Atlas 200 DK开发者套件",{"type":17,"tag":25,"props":130,"children":131},{},[132],{"type":23,"value":133},"Atlas 200 DK开发者套件（型号:3000）是一款高性能的AI应用开发板，集成了昇腾处理器，提供强大的AI计算能力，支持快速开发和验证AI应用，适用于开发者方案验证、高校教育、科学研究等场景。",{"type":17,"tag":25,"props":135,"children":136},{},[137],{"type":23,"value":138},"集群服务器可由树莓派等Linux主机系统构建，用于接入多个Atlas 200 DK开发者套件，为深度学习实验提供充足的AI算力池。",{"type":17,"tag":25,"props":140,"children":141},{},[142],{"type":23,"value":143},"（2）软件平台：开发集群服务器软件，提供高效的并行推理服务，支持远程开展深度学习的推理实验，主要包括：集群管理系统，提供简单、灵活的模型部署等功能；远程推理软件包，支持Python等多种开发语言，适用于不同课程，匹配学生的个性化选择。",{"type":17,"tag":25,"props":145,"children":146},{},[147],{"type":23,"value":148},"该集群实验平台具有显著优势。使用“昇腾AI远程推理集群”进行推理实验，无需人手一套Atlas 200 DK开发套件，无需物理连接，无需登录使用。使用少量的AI算力设备即可开展大量学生的实验课，这极大减少了对AI算力设备的数量需求，并且大幅降低了实验准备的复杂性，让学生专注深度学习的实验本身，提升实验教学效果。",{"type":17,"tag":25,"props":150,"children":151},{},[152],{"type":17,"tag":64,"props":153,"children":155},{"alt":7,"src":154},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2024/08/23/535465cfbe5d4b4e8228fef08f441122.png",[],{"type":17,"tag":25,"props":157,"children":158},{},[159],{"type":23,"value":160},"图3. 昇腾AI远程推理集群的使用模式",{"type":17,"tag":25,"props":162,"children":163},{},[164],{"type":17,"tag":29,"props":165,"children":166},{},[167],{"type":23,"value":168},"02",{"type":17,"tag":25,"props":170,"children":171},{},[172],{"type":17,"tag":29,"props":173,"children":174},{},[175],{"type":23,"value":176},"实验项目建设",{"type":17,"tag":25,"props":178,"children":179},{},[180],{"type":23,"value":181},"整个项目涉及深度学习的数据集构造、模型训练、模型推理和应用开发的全流程内容，一方面将帮助学生掌握深度学习应用的全部关键环节，敢于独立开展创新实践，另一方面将帮助学生了解人工智能领域的先进产业技术，紧跟行业发展前沿，培养创新意识。",{"type":17,"tag":25,"props":183,"children":184},{},[185],{"type":23,"value":186},"深度学习应用的基本开发流程如图4所示，本实验项目覆盖其中的全部内容，培养人工智能系统的设计和开发的实践与创新能力。",{"type":17,"tag":25,"props":188,"children":189},{},[190],{"type":17,"tag":64,"props":191,"children":193},{"alt":7,"src":192},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2024/08/23/407f013c07f046b2b2a8e3675718cba3.png",[],{"type":17,"tag":25,"props":195,"children":196},{},[197],{"type":23,"value":198},"图4. 深度学习应用的基本开发流程",{"type":17,"tag":25,"props":200,"children":201},{},[202],{"type":23,"value":203},"本实验项目包括3个实验，分别是：LeNet-5的昇思MindSpore框架训练、基于Atlas 200 DK的手写数字识别和ModelArts深度学习模型训练与部署。这三个实验环环相扣，逐级挑战，覆盖深度学习应用的全流程，即：数据准备、模型训练、模型评估和模型部署，培养人工智能领域的综合实践创新能力。第一个项目独立完成，第二个项目和第三个项目以小组为单位完成，每组2～3人，锻炼学生对AI项目的管理能力、沟通交流能力以及团队协作能力。",{"type":17,"tag":25,"props":205,"children":206},{},[207],{"type":23,"value":208},"以下简要介绍这三个实验项目。",{"type":17,"tag":25,"props":210,"children":211},{},[212],{"type":17,"tag":29,"props":213,"children":214},{},[215],{"type":23,"value":216},"（1） 项目1: LeNet-5的MindSpore框架训练",{"type":17,"tag":25,"props":218,"children":219},{},[220],{"type":23,"value":221},"模型训练是深度学习应用的前提，本实验使用昇思MindSpore框架在通用CPU上训练LeNet-5模型，帮助学生深刻理解模型训练的原理，掌握使用华为AI技术的深度学习模型训练方法。",{"type":17,"tag":25,"props":223,"children":224},{},[225],{"type":23,"value":226},"LeNet-5是深度学习的经典网络模型，在手写数字识别任务上取得了显著效果，并获广泛应用，并为后来卷积神经网络的发展奠定了重要基础。LeNet-5主要由两个卷积层和三个全连接层组成，其中卷积层用于提取图像特征，全连接层用于对提取的特征进行分类。",{"type":17,"tag":25,"props":228,"children":229},{},[230],{"type":17,"tag":64,"props":231,"children":233},{"alt":7,"src":232},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2024/08/23/adf86cc823c845628337e9de06d47c91.png",[],{"type":17,"tag":25,"props":235,"children":236},{},[237],{"type":23,"value":238},"图5. LeNet-5模型的神经网络结构",{"type":17,"tag":25,"props":240,"children":241},{},[242],{"type":23,"value":243},"昇思MindSpore是华为推出的新一代全场景AI框架，作为AI基础软件，原生支持大模型的训练与推理、AI+科学计算创新，最佳亲和昇腾硬件释放极致性能。主要的3个特点：",{"type":17,"tag":25,"props":245,"children":246},{},[247],{"type":23,"value":248},"1）极简易用，一套架构支持动静统一，兼顾大模型训练效率与性能；",{"type":17,"tag":25,"props":250,"children":251},{},[252],{"type":23,"value":253},"2）训推一体，模型从训练到推理部署无需切换框架、0代码转换；3提供开箱即用的典型模型如Llama、InternLM2、Qwen等及大模型开发套件，使能开发者加速大模型创新。",{"type":17,"tag":25,"props":255,"children":256},{},[257],{"type":23,"value":258},"昇思MindSpore作为框架后起之秀，自2020年开源以来锐意进取，已成为国内第一梯队框架，技术与开源生态繁荣，版本累计下载次数超过750万，核心贡献者超过3.2万人，支持国内外50多个大模型，广泛应用于金融、运营商、制造、教育等10多个行业，展现出强大的技术实力和应用前景。",{"type":17,"tag":25,"props":260,"children":261},{},[262],{"type":23,"value":263},"本实验要求学生掌握以下主要内容：",{"type":17,"tag":25,"props":265,"children":266},{},[267],{"type":23,"value":268},"1）使用昇思MindSpore定义LeNet-5深度神经网络；",{"type":17,"tag":25,"props":270,"children":271},{},[272],{"type":23,"value":273},"2）使用昇思MindSpore定义损失函数和优化器；",{"type":17,"tag":25,"props":275,"children":276},{},[277],{"type":23,"value":278},"3）使用昇思MindSpore加载Mnist数据集；",{"type":17,"tag":25,"props":280,"children":281},{},[282],{"type":23,"value":283},"4）使用CPU进行训练，并评估模型性能；",{"type":17,"tag":25,"props":285,"children":286},{},[287],{"type":23,"value":288},"5）查看结果，并保存模型文件；",{"type":17,"tag":25,"props":290,"children":291},{},[292],{"type":23,"value":293},"6）创新探索：鼓励学生探讨LeNet-5的卷积核大小、网络结构等模型超参数可如何调整，并分析这些调整将带来哪些影响。",{"type":17,"tag":25,"props":295,"children":296},{},[297],{"type":23,"value":298},"该实验的LeNet-5模型文件将用于项目2的模型推理和应用。",{"type":17,"tag":25,"props":300,"children":301},{},[302],{"type":17,"tag":29,"props":303,"children":304},{},[305],{"type":23,"value":306},"（2）项目2：基于Atlas 200 的手写数字识别",{"type":17,"tag":25,"props":308,"children":309},{},[310],{"type":23,"value":311},"深度学习模型的推理是深度学习应用的基础，本实验以手写数字识别应用为例，使用华为Atlas 200I A2推理卡集群或Atlas 200I DK A2作为模型部署和推理设备，帮助学生深入理解模型推理的原理，掌握使用华为AI技术的深度学习模型推理方法。",{"type":17,"tag":25,"props":313,"children":314},{},[315],{"type":23,"value":316},"手写数字识别是CNN（卷积神经网络）模型的第一个成功应用，被广泛应用于银行支票手写数字识别、卡号识别等多种实际场景，是人工智能领域的一个经典实验项目。",{"type":17,"tag":25,"props":318,"children":319},{},[320],{"type":23,"value":321},"本实验使用项目1的实验结果，在“昇腾AI远程推理集群”上进行推理。实验可在台式计算机或笔记本电脑上进行，要求掌握以下主要内容：",{"type":17,"tag":25,"props":323,"children":324},{},[325],{"type":23,"value":326},"1）使用模型工具atc转换项目1的模型文件为Atlas 200 DK使用的离线模型格式；",{"type":17,"tag":25,"props":328,"children":329},{},[330],{"type":23,"value":331},"2）将上步得到的离线模型上传到Atlas 200 DK套件上，并使用远程推理工具airloader加载为远程推理服务；",{"type":17,"tag":25,"props":333,"children":334},{},[335],{"type":23,"value":336},"3）安装远程推理软件包，下载远程推理的样例开源项目，查看手写数字识别的Python文件；",{"type":17,"tag":25,"props":338,"children":339},{},[340],{"type":23,"value":341},"4）使用该远程推理样例项目，对输入的图片进行数字识别，分析结果；",{"type":17,"tag":25,"props":343,"children":344},{},[345],{"type":23,"value":346},"5）创新探索：鼓励学生参照该样例，尝试其他模型完成推理部署和创新应用开发，鼓励学生深入了解Atlas 200I A2模块，并尝试基于此智能硬件开展更多创新探索。",{"type":17,"tag":25,"props":348,"children":349},{},[350],{"type":23,"value":351},"该实验帮助学生掌握了深度学习的推理流程和应用方法，由于使用“昇腾AI远程推理集群”进行，极大地降低了深度学习平台的建设成本和使用复杂性，有利于该实验的推广应用，服务于更多课程的教学需要。",{"type":17,"tag":25,"props":353,"children":354},{},[355],{"type":17,"tag":29,"props":356,"children":357},{},[358],{"type":23,"value":359},"（3） 项目3：云侧：基于ModelArts的深度学习模型训练与部署",{"type":17,"tag":25,"props":361,"children":362},{},[363],{"type":23,"value":364},"在掌握了端侧模型训练与模型推理的基础上，本实验在华为云上使用ModelArts 和昇思MindSpore作为实验环境，覆盖数据集准备、CNN的模型结构设计、模型训练和模型部署的完整流程。",{"type":17,"tag":25,"props":366,"children":367},{},[368],{"type":23,"value":369},"ModelArts是华为云推出的一款面向开发者的一站式AI开发平台，旨在提供简单、高效、全面的AI开发解决方案。该平台集成了丰富的功能和工具，支持海量数据预处理及交互式智能标注、大规模分布式训练、自动化模型生成，以及端-边-云模型按需部署能力。",{"type":17,"tag":25,"props":371,"children":372},{},[373],{"type":23,"value":374},"本实验在华为云上进行，使用华为全栈AI技术开发基于深度学习技术的花卉识别网络服务，主要任务包括：准备数据集、构建CNN模型、训练模型和部署为网络服务，要求学生掌握以下主要内容：",{"type":17,"tag":25,"props":376,"children":377},{},[378],{"type":23,"value":379},"1）创建ModelArts实验环境，导入昇思MindSpore等相关软件包；",{"type":17,"tag":25,"props":381,"children":382},{},[383],{"type":23,"value":384},"2）导入花卉数据集，构建训练集和测试集，并进行可视化分析；",{"type":17,"tag":25,"props":386,"children":387},{},[388],{"type":23,"value":389},"3）基于昇思MindSpore构建ResNet等深度CNN神经网络，定义网络结构和优化器；",{"type":17,"tag":25,"props":391,"children":392},{},[393],{"type":23,"value":394},"4）使用NPU进行模型训练，分析训练过程和收敛结果；",{"type":17,"tag":25,"props":396,"children":397},{},[398],{"type":23,"value":399},"5）使用测试数据集进行模型验证，评估模型性能；",{"type":17,"tag":25,"props":401,"children":402},{},[403],{"type":23,"value":404},"6）在华为云上部署为花卉识别网络服务，验证并分析应用性能；",{"type":17,"tag":25,"props":406,"children":407},{},[408],{"type":23,"value":409},"7）创新探索：鼓励学生基于华为技术开发GAN、RNN等其它深度神经网络模型的创新AI应用，并积极参与科研实践和科创竞赛，进一步锻炼创新和综合实践能力。",{"type":17,"tag":25,"props":411,"children":412},{},[413],{"type":23,"value":414},"该实验内容涵盖深度学习应用开发的全流程，帮助学生掌握华为AI技术使用和人工智能应用开发技能，且全部操作使用华为云服务进行，实验成本低。",{"type":17,"tag":25,"props":416,"children":417},{},[418],{"type":17,"tag":29,"props":419,"children":420},{},[421],{"type":23,"value":422},"项目成果、创新点及成效",{"type":17,"tag":25,"props":424,"children":425},{},[426],{"type":23,"value":427},"项目获“优秀”结题，并获得教育部产学合作协同育人项目华为优秀成果奖。",{"type":17,"tag":25,"props":429,"children":430},{},[431],{"type":17,"tag":64,"props":432,"children":434},{"alt":7,"src":433},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2024/08/23/c366ab73dc3a4e86ab957ed3c8aa4d0b.png",[],{"type":17,"tag":25,"props":436,"children":437},{},[438],{"type":23,"value":439},"图6. 项目所获荣誉证书",{"type":17,"tag":25,"props":441,"children":442},{},[443],{"type":17,"tag":29,"props":444,"children":445},{},[446],{"type":23,"value":447},"0****1",{"type":17,"tag":25,"props":449,"children":450},{},[451],{"type":17,"tag":29,"props":452,"children":453},{},[454],{"type":23,"value":455},"项目成果",{"type":17,"tag":25,"props":457,"children":458},{},[459],{"type":17,"tag":29,"props":460,"children":461},{},[462],{"type":23,"value":463},"（1）“昇腾AI推理集群”实践新平台",{"type":17,"tag":25,"props":465,"children":466},{},[467],{"type":23,"value":468},"针对传统的人工智能实践模式的资源依赖强、平台高成本、实践项目不灵活等不足，本项目独立开发了“昇腾AI推理集群实践平台”实验新方案，解决了Atlas 200 DK开发板资源不足、环境配置繁琐、实践结果不直观等痛点问题，基于此构建了一种分布式人工智能教学实践平台，符合高校实践教学迫切需要。",{"type":17,"tag":25,"props":470,"children":471},{},[472],{"type":17,"tag":64,"props":473,"children":475},{"alt":7,"src":474},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2024/08/23/447b0dde04cf429aba68d042bee4ee5e.png",[],{"type":17,"tag":25,"props":477,"children":478},{},[479],{"type":23,"value":480},"图7. “昇腾AI远程推理集群”实物图",{"type":17,"tag":25,"props":482,"children":483},{},[484],{"type":17,"tag":29,"props":485,"children":486},{},[487],{"type":23,"value":488},"（2）共建《机器学习》课程，有机融入昇思MindSpore等华为AI全栈技术的课程教学资源",{"type":17,"tag":25,"props":490,"children":491},{},[492],{"type":23,"value":493},"主要包括：",{"type":17,"tag":495,"props":496,"children":497},"ul",{},[498,504,509,514,519],{"type":17,"tag":499,"props":500,"children":501},"li",{},[502],{"type":23,"value":503},"《机器学习》的课程教案、课件；",{"type":17,"tag":499,"props":505,"children":506},{},[507],{"type":23,"value":508},"实践项目方案和实验手册；",{"type":17,"tag":499,"props":510,"children":511},{},[512],{"type":23,"value":513},"学生实践项目报告和成果DEMO；",{"type":17,"tag":499,"props":515,"children":516},{},[517],{"type":23,"value":518},"相关完整的数据集和源代码；",{"type":17,"tag":499,"props":520,"children":521},{},[522],{"type":23,"value":523},"使用心得反馈和建议。",{"type":17,"tag":25,"props":525,"children":526},{},[527],{"type":17,"tag":29,"props":528,"children":529},{},[530],{"type":23,"value":531},"（3）昇腾AI远程推理集群开源项目",{"type":17,"tag":25,"props":533,"children":534},{},[535],{"type":23,"value":536},"已在gitee网站上公开发布，访问IP地址超过3900个。共享内容包括Atlas 200 DK远程推理镜像、相关软件包、实验案例源代码和实验手册等。",{"type":17,"tag":25,"props":538,"children":539},{},[540],{"type":17,"tag":29,"props":541,"children":542},{},[543],{"type":23,"value":168},{"type":17,"tag":25,"props":545,"children":546},{},[547],{"type":17,"tag":29,"props":548,"children":549},{},[550],{"type":23,"value":551},"项目创新点",{"type":17,"tag":25,"props":553,"children":554},{},[555],{"type":23,"value":556},"华为公司自主研发的昇腾全栈AI平台，具有强大算力和生态技术先进性，并通过Atlas 200 DK系列推理设备，为开展深度学习课程实践和人才培养提供了极好的条件，但存在设备资源稀缺和连接复杂性等困境。",{"type":17,"tag":25,"props":558,"children":559},{},[560],{"type":23,"value":561},"本项目采用将深度神经网络模型的推理与前后处理相分离的新思路，创新提出并独立完成的“Atlas 200 DK远程推理集群”AI实验新方案，支持分布式并行推理、24小时远程使用，可解决硬件资源稀缺、环境配置复杂等难点问题，具有低成本、伸缩性强、易于上手、可视化体验好的特点，对提升人工智能教学的深度、难度和强度具有重要推广价值。",{"type":17,"tag":25,"props":563,"children":564},{},[565],{"type":23,"value":566},"主要创新点:",{"type":17,"tag":25,"props":568,"children":569},{},[570],{"type":23,"value":571},"1）解除开发板物理连接，实现远程推理服务，简化部署，全时可用。解决了神经网络计算设备严重不足、深度学习环境配置复杂、上手难的问题；",{"type":17,"tag":25,"props":573,"children":574},{},[575],{"type":23,"value":576},"2）提供跨平台推理框架，自动分布运行，学生专注模型应用开发。该框架将模型推理与深度学习应用相分离，可充分发挥华为AI设备的NPU算力和学生个人电脑的CPU算力，支持开展趣味强、样式多的创新实验；",{"type":17,"tag":25,"props":578,"children":579},{},[580],{"type":23,"value":581},"3）提供实验新案例和测试工具。基于该实验平台，设计开发了图像风格转换、人像肖像分割等远程推理新案例；开发了灵活、一致的测试工具，提升实验体验；",{"type":17,"tag":25,"props":583,"children":584},{},[585],{"type":23,"value":586},"4）提供集群解决方案，进一步充分发挥昇腾算力，服务更多学生。构建的集群服务提供透明的分布式服务，可提供并行推理能力，服务更多学生在少量开发板资源上同时开展实践。",{"type":17,"tag":25,"props":588,"children":589},{},[590],{"type":17,"tag":29,"props":591,"children":592},{},[593],{"type":23,"value":594},"5）从训练到推理全流程使用华为全栈AI技术，包括昇腾芯片、CANN、昇思MindSpore、ModelArts，可在华为云或昇腾训练服务器上训练模型，在开发板上推理，完成更多个性化的创新应用。",{"type":17,"tag":25,"props":596,"children":597},{},[598],{"type":17,"tag":29,"props":599,"children":600},{},[601],{"type":23,"value":602},"03",{"type":17,"tag":25,"props":604,"children":605},{},[606],{"type":17,"tag":29,"props":607,"children":608},{},[609],{"type":23,"value":610},"项目成效",{"type":17,"tag":25,"props":612,"children":613},{},[614],{"type":23,"value":615},"自2020年秋学期始，电子科技大学自动化工程学院逐步全面引入昇腾与昇思MindSpore的国产软硬件体系，并作为首批“智能基座院校”大力推进“智能基座”课程建设。截止2024年春季学期，已达成全学院AI类课程与昇思MindSpore全面融合覆盖。如下表：",{"type":17,"tag":25,"props":617,"children":618},{},[619],{"type":17,"tag":64,"props":620,"children":622},{"alt":7,"src":621},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2024/08/23/d3ccb3d547a840e384068d1992b44aba.png",[],{"type":17,"tag":25,"props":624,"children":625},{},[626],{"type":23,"value":627},"经过2年有余的课程融合，使用昇思MindSpore授课累计覆盖学生500+。基于此建成国内首个昇腾AI推理集群实践平台，已应用于12次教学任务，其中本科生课程6次，研究生课程6次；受益学生累计超过860人，总体效果明显。",{"type":17,"tag":25,"props":629,"children":630},{},[631],{"type":17,"tag":64,"props":632,"children":634},{"alt":7,"src":633},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2024/08/23/4cc58f722431440ea62c10dd5c2602aa.png",[],{"type":17,"tag":25,"props":636,"children":637},{},[638],{"type":23,"value":639},"图8. 课堂实验场景",{"type":17,"tag":25,"props":641,"children":642},{},[643],{"type":17,"tag":64,"props":644,"children":646},{"alt":7,"src":645},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2024/08/23/d4b54a0b71184250a07e1fe01185051f.png",[],{"type":17,"tag":25,"props":648,"children":649},{},[650],{"type":23,"value":651},"图9. 实验成果展示",{"type":17,"tag":25,"props":653,"children":654},{},[655],{"type":23,"value":656},"学生实验成果丰富，评教结果良好。研究生教学任务已完成6次，累计612人选课，评教平均分91.2。本科生教学任务已完成6次，累计参评人数224，学生评教均为“四星”。",{"type":17,"tag":25,"props":658,"children":659},{},[660],{"type":17,"tag":64,"props":661,"children":663},{"alt":7,"src":662},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2024/08/23/8c53efca5c2542ce853a9c6c5b4fb143.png",[],{"type":17,"tag":25,"props":665,"children":666},{},[667],{"type":23,"value":668},"图10. 学生反馈（节选）",{"type":17,"tag":25,"props":670,"children":671},{},[672],{"type":23,"value":673},"指导学生参与昇腾AI相关竞赛获奖超过10次，主要奖项有：",{"type":17,"tag":495,"props":675,"children":676},{},[677,682,687,692],{"type":17,"tag":499,"props":678,"children":679},{},[680],{"type":23,"value":681},"昇腾AI创新大赛，四川省银奖2项，2022，2023",{"type":17,"tag":499,"props":683,"children":684},{},[685],{"type":23,"value":686},"昇腾AI创新大赛，四川省新锐奖2项，2022，2023",{"type":17,"tag":499,"props":688,"children":689},{},[690],{"type":23,"value":691},"华为ICT大赛实践赛AI赛道，全国一等奖1项，2024",{"type":17,"tag":499,"props":693,"children":694},{},[695],{"type":23,"value":696},"华为ICT大赛实践赛AI赛道，全球二等奖1项，2024",{"type":17,"tag":25,"props":698,"children":699},{},[700],{"type":17,"tag":29,"props":701,"children":702},{},[703],{"type":23,"value":704},"应用及推广情况",{"type":17,"tag":25,"props":706,"children":707},{},[708],{"type":23,"value":709},"面向全国合作教师、校内外学生社团/开发者等进行15次经验分享，均获得高度评价，培训视频发布在各官方网站。",{"type":17,"tag":25,"props":711,"children":712},{},[713],{"type":23,"value":714},"此外，在华为云开发者社区论坛发帖，浏览量超过9500次。",{"type":17,"tag":25,"props":716,"children":717},{},[718],{"type":17,"tag":29,"props":719,"children":720},{},[721],{"type":23,"value":722},"经验总结",{"type":17,"tag":25,"props":724,"children":725},{},[726],{"type":23,"value":727},"项目的主要经验有：",{"type":17,"tag":25,"props":729,"children":730},{},[731],{"type":23,"value":732},"1）产业前沿技术是培养卓越工程师型人才的重要实践载体。昇腾AI处理器和昇思MindSpore AI框架一站式支持AI相关课程的全流程实践，赋能自主工业软件技术先进、资源丰富、支持全面；openEuler新型OS是自主基础软件的优秀实践平台。",{"type":17,"tag":25,"props":734,"children":735},{},[736],{"type":23,"value":737},"2）产教融合需要校企结合、主动融合。根据课程需要，主动了解企业先进资源，并针对性对接企业资源、融合企业资源。如不能精准对接，则需要先“融”才能“合”，重构课程或开发新工具。",{"type":17,"tag":25,"props":739,"children":740},{},[741],{"type":23,"value":742},"3）昇腾AI推理集群平台实践新方案是产教融合的典型案例：与企业深入沟通，通过建设新工具，实现了充分融合，可低成本地支持深度学习课程的实践教学，有效培养学生的创新实践能力和创新意识。",{"type":17,"tag":25,"props":744,"children":745},{},[746],{"type":23,"value":747},"4）大模型引发产业巨变，产教融合是课程改革新方向、新机会，深化产学合作，建设跟踪前沿产业技术的课程实践资源，服务更多学生。",{"type":17,"tag":25,"props":749,"children":750},{},[751],{"type":23,"value":752},"项目实施以来，课程组在持续开展教学实践，并不断深入探索优化，激发了学生的昇腾实践兴趣，取得了学生的广泛认可。",{"title":7,"searchDepth":754,"depth":754,"links":755},4,[],"markdown","content:news:zh:3271.md","content","news/zh/3271.md","news/zh/3271","md",1776506082591]