[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"site-home-live-entry-setting":3,"site-header-contests":8,"site-header-camps":49,"study-material-detail-study-11":104},{"enabled":4,"title":5,"statusText":5,"buttonText":5,"campSlug":5,"stage":6,"classroomId":7,"hasClassroom":4},false,"",0,null,{"total":9,"list":10,"pageNum":44,"pageSize":45,"size":9,"startRow":44,"endRow":9,"pages":44,"prePage":6,"nextPage":6,"isFirstPage":46,"isLastPage":46,"hasPreviousPage":4,"hasNextPage":4,"navigatePages":47,"navigatepageNums":48,"navigateFirstPage":44,"navigateLastPage":44},3,[11,25,35],{"slug":12,"title":13,"logo":7,"liveUrl":14,"prize":15,"sponsor":16,"description":17,"organizer":18,"startTime":19,"endTime":20,"registrationDeadline":21,"status":22,"dashboardVisibility":23,"createdAt":24},"spring2026","2026春季启元人工智能大赛","https:\u002F\u002Fwww.bilibili.com\u002Fblackboard\u002Flive\u002Flive-activity-player.html?cid=1822703816&quality=0","200,000","麒麟软件有限公司","“2026春季启元人工智能大赛”是由启元实验室主办的面向全国人工智能领域开发者、科研人员及高校学生的权威赛事。大赛旨在促进与培养具备工程实践能力和创新能力的优秀人工智能领域开发者、推广国产软硬件技术与生态、助力国家人工智能产业高质量发展。\n各参赛选手需在规定的比赛周期内，依据赛事规则选定赛题并完成相应挑战，优胜者将共同角逐总额20万元的赛事奖金池，并且还将有机会获得实习和知名企业优质岗位内推资格。","启元实验室","2026-04-19 16:00:00","2026-07-13 16:00:00","2026-06-16 16:00:00","ongoing","private","2026-04-16 08:00:00",{"slug":26,"title":27,"logo":7,"liveUrl":14,"prize":28,"sponsor":7,"description":29,"organizer":18,"startTime":30,"endTime":31,"registrationDeadline":32,"status":33,"dashboardVisibility":23,"createdAt":34},"autumn2025","2025秋季启元人工智能大赛","150,000","“2025秋季启元人工智能大赛”是由启元实验室主办的面向全国人工智能领域开发者、科研人员及高校学生的权威赛事。大赛旨在促进与培养具备工程实践能力和创新能力的优秀人工智能领域开发者、推广国产软硬件技术与生态、助力国家人工智能产业高质量发展。","2025-11-08 08:00:00","2026-01-13 08:00:00","2025-12-30 08:00:00","ended","2025-11-03 08:00:00",{"slug":36,"title":37,"logo":7,"liveUrl":14,"prize":38,"sponsor":7,"description":39,"organizer":18,"startTime":40,"endTime":41,"registrationDeadline":42,"status":33,"dashboardVisibility":23,"createdAt":43},"summer2025","启元人工智能大赛(系统方向)","80,000","启元人工智能大赛由启元实验室主办，其中系统赛道基于启元实验室九源软件栈的相关开源项目，通过拉动开源社区力量，一方面促进国产智能计算生态的发展，另一方面也为人工智能系统方向的学生及从业者提供一个实践平台。","2025-07-08 08:00:00","2025-08-25 08:00:00","2025-07-22 08:00:00","2025-06-22 08:00:00",1,5,true,8,[44],{"total":45,"list":50,"pageNum":44,"pageSize":45,"size":45,"startRow":44,"endRow":45,"pages":44,"prePage":6,"nextPage":6,"isFirstPage":46,"isLastPage":46,"hasPreviousPage":4,"hasNextPage":4,"navigatePages":47,"navigatepageNums":103,"navigateFirstPage":44,"navigateLastPage":44},[51,62,73,84,94],{"id":52,"slug":53,"title":54,"logo":55,"publicLiveEmbedUrl":14,"description":56,"introduction":57,"organizer":18,"coOrganizer":5,"sponsor":16,"startTime":58,"endTime":59,"status":22,"dashboardVisibility":60,"createdAt":58,"updatedAt":61},"019eeeff-6b00-7d55-a629-efafded3e232","summer2026","2026年夏季InfiniTensor大模型与人工智能系统训练营","https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fuploads\u002F2026-06\u002F019eef7f-7816-7828-9309-79c2c61dd568.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=DdQTMKVo1nful94EsDeKYRm5DII%3D","由启元实验室发起的InfiniTensor大模型与人工智能系统训练营，面向开源社区、高校及产业界伙伴开放。活动以「普及AI全栈知识技术、培养工程实践与创新型人才、推广国产软硬件技术生态」为核心目标，助力国家人工智能产业高质量发展。","...","2026-06-22 08:00:00","2026-10-19 08:00:00","public","2026-10-12 08:00:00",{"id":63,"slug":64,"title":65,"logo":66,"publicLiveEmbedUrl":7,"description":67,"introduction":57,"organizer":18,"coOrganizer":5,"sponsor":68,"startTime":69,"endTime":70,"status":33,"dashboardVisibility":23,"createdAt":71,"updatedAt":72},"019af3f3-e06a-7245-aa8d-37dc99a69c73","winter2025","2025年冬季InfiniTensor大模型与人工智能系统训练营","https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fuploads\u002F2025-12\u002F019afc69-b7ca-7805-91fb-605429671c01.png?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=HWdVGs1NgNzaJvqpNa66tZqqOTo%3D","由启元实验室发起的InfiniTensor大模型与人工智能系统训练营，面向开源社区、高校及产业界伙伴开放。活动以「普及AI全栈知识技术、培养工程实践与创新型人才、推广国产软硬件技术生态」为核心目标，助力国家人工智能产业高质量发展。 \n","新华三技术有限公司","2025-12-25 08:00:00","2026-03-27 08:00:00","2025-12-06 08:00:00","2026-10-22 20:24:30",{"id":74,"slug":36,"title":75,"logo":76,"publicLiveEmbedUrl":7,"description":77,"introduction":57,"organizer":18,"coOrganizer":5,"sponsor":78,"startTime":79,"endTime":80,"status":81,"dashboardVisibility":23,"createdAt":82,"updatedAt":83},"01980342-208c-7803-b3f3-19e1d31960ba","2025年夏季InfiniTensor大模型与人工智能系统训练营","https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fuploads\u002F2025-07\u002F0198045e-28fe-78f0-95fe-23835ce58395.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=5IITjeecr8z%2B54QsS2SaoytOVRM%3D","InfiniTensor大模型与人工智能系统训练营由启元实验室及清华大学 AI 系统方向团队发起，面向开源社区、高校和产业界，普及人工智能各方面知识和技术，培养相关人才。训练营联合开源项目开发者、学术界大牛、工业界资深工程师参与授课与项目指导。课程采用「线上直播授课 + 在线答疑 + 项目实践辅导 」的方式进行，确保每位同学都能获得充分成长。训练营全程免费，通过视频、直播课、习题和项目实践，在人工智能飞速发展的背景下，为学员提供一窥深层技术的机会。","北京积算科技有限公司","2025-07-14 08:00:00","2025-09-30 08:00:00","upcoming","2025-07-13 08:00:00","2026-10-22 20:24:58",{"id":85,"slug":86,"title":87,"logo":88,"publicLiveEmbedUrl":7,"description":89,"introduction":57,"organizer":18,"coOrganizer":5,"sponsor":5,"startTime":90,"endTime":91,"status":81,"dashboardVisibility":23,"createdAt":92,"updatedAt":93},"01980cf9-a7d8-7796-ae86-0bbe05334add","winter2024","2024年冬季InfiniTensor大模型与人工智能系统训练营","https:\u002F\u002Ftesting-cdn.infinitensor.com\u002Fuploads\u002F2025-07\u002F01981269-ce23-7624-a287-9f07dd0d85f6.jpg","InfiniTensor大模型与人工智能系统训练营由清华大学发起，面向开源社区、高校和产业界，普及人工智能各方面知识和技术，培养相关人才。训练营全程免费，通过视频、直播课、习题和项目实践，在人工智能飞速发展的背景下，为学员提供一窥深层技术的机会。报名学员请阅读“导学”课程页的训练营介绍。","2025-01-05 08:00:00","2025-03-22 08:00:00","2025-07-15 08:00:00","2026-10-22 20:24:39",{"id":95,"slug":96,"title":97,"logo":98,"publicLiveEmbedUrl":7,"description":99,"introduction":57,"organizer":18,"coOrganizer":5,"sponsor":5,"startTime":100,"endTime":101,"status":81,"dashboardVisibility":23,"createdAt":92,"updatedAt":102},"01980c20-363a-7dc3-8f5a-58a68e08dcc4","summer2024","2024年夏季InfiniTensor大模型与人工智能系统训练营","https:\u002F\u002Ftesting-cdn.infinitensor.com\u002Fuploads\u002F2025-07\u002F01981269-fbfa-788a-afcd-17fa9be74ff9.jpg","InfiniTensor大模型与人工智能系统训练营由清华大学发起，面向开源社区和高校，普及人工智能各方面知识和技术，培养相关人才。训练营全程免费，通过课程、习题和项目实践，在人工智能飞速发展的背景下，为学员提供一窥深层技术的机会。","2024-07-20 08:00:00","2024-09-29 08:00:00","2026-10-22 20:24:49",[44],{"currentItem":105,"processedData":125},{"img":106,"keywords":107,"treeType":44,"description":108,"showStatus":44,"updateTime":109,"type":110,"title":111,"content":112,"nextArticle":113,"createTime":109,"sortOrder":117,"canonicalPath":118,"id":119,"contentType":120,"prevArticle":121},"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_11_fm.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=2CGGxx1zjCZ1XWH8%2FAgd4pLs2TE%3D","InfiniTensor,冬季训练营,训练系统,训练流程,工程基础","讲解训练系统基础介绍，涵盖训练流程、关键组件与训练工程入门知识。","2026-10-22 20:38:11",2,"InfiniTensor 冬季训练营训练系统第一课《训练基础介绍》","![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_11_zw01.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=YIJ4V771ULsBRKLQqctuduXY3tc%3D)\r\n\r\n\r\n2026 年 1 月 13 日，2025 冬季 InfiniTensor 大模型与人工智能系统训练营训练系统系列第一课正式开讲，课程主题为《训练基础介绍》。\r\n\r\n本节课程主要介绍一些训练相关的基础知识，同时介绍一下团队正在开发的自研训练框架 InfiniTrain。\r\n\r\n### 神经网络基础概念\r\n神经网络的基本组成结构\r\n- 输入层 \u002F 隐藏层 \u002F 输出层\r\n\r\n\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_11_zw02.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=ZmYszhJUzPfNt4uBBZ%2Bysni2FhI%3D)\r\n\r\n\r\n\r\n### 优化算法与损失函数\r\n#### 1.梯度下降法\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_11_zw03.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=u6swUoVx4fEzHj%2F%2FSj3XcBnRloo%3D)\r\n\r\n\r\n#### 2.损失函数类型\r\n交叉熵损失（分类任务）\r\n均方误差损失（回归任务）\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_11_zw04.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=wXAxMmLj%2B8%2F83StgNmmm4jzLJe0%3D)\r\n\r\n### 激活函数详解\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_11_zw05.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=xmitMfGBaBeQeWOv02qbYcWmzYA%3D)\r\n\r\n### 反向传播的教学基础\r\n- 链式法则数学原理\r\n- 反向传播具体流程\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_11_zw06.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=VvHXK2QHwYUZ%2Ft%2FWI3mRTmvtRmU%3D)\r\n\r\n\r\n\r\n### PyTorch 核心机制\r\n#### 1.Autograd 自动微分机制\r\n#### 2.构建 Function 计算图\r\n\r\n\r\n#### 3.PyTorch 梯度计算规则\r\n#### 4.requires_grad 参数作用\r\n\r\n\r\n### 混合精度训练\r\n- torch.amp\r\n- Nvidia APEX\r\n- 分层混合精度策略\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_11_zw07.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=iFV%2FPRx2Q7DJZp7Sdf3%2BYcWW0mo%3D)\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_11_zw08.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=vJtJEQ2xu%2FL42ajqaGVcuvm80Qo%3D)\r\n\r\n\r\n### 分布式训练策略\r\n#### 1.并行训练基础\r\n- 数据并行原理\r\n- 模型并行策略\r\n- 混合并行方法\r\n#### 2.高级并行技术\r\n- 张量并行实现\r\n- 序列并行优化\r\n- 流水线并行设计\r\n\r\n\r\n### 显存优化技巧\r\n#### 1.大模型训练挑战\r\n- 显存占用分析\r\n- 优化策略概述\r\n#### 2.具体优化技术\r\n- ZeRO (Zero Redundancy Optimizer)\r\n- 梯度检查点机制\r\n- 显存高效训练方法\r\n\r\n\r\n### 强化学习与 DeepSeek 入门\r\n#### 1.强化学习基础\r\n- 基本概念与原理\r\n- 常用算法介绍\r\n#### 2.DeepSeek\r\n- 架构特点与应用\r\n- 实践入门指导\r\n- 实践环节与作业要求\r\n\r\n### 自研框架 InfiniTrain\r\n- 框架架构介绍\r\n- Kernel 层（CUDA 内核）\r\n- Autograd 层（自动微分）\r\n- NN层（模型构建）\r\n- 示例模型实现\r\n- GPT-2 训练示例\r\n- LLaMA-3 训练示例\r\n\r\n### 总结与展望\r\n本节课深入理解前向传播、反向传播与梯度优化等训练核心基础。后续将系统学习分布式策略与自研框架实践，全面提升工程能力。\r\n\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_01_zw6.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=Jvvk9S3asMdwJQP5eMwtAGsqW0k%3D)\r\n\r\n\r\n### 💡 学习贴士\r\n更多细节及完整课程内容，请观看直播或查看课程回放。\r\n\r\n链接：https:\u002F\u002Fwww.infinitensor.com\u002Fcamp\u002Fwinter2025\r\n\r\n### 📚 训练营报名持续进行中\r\n扫码报名与加入群聊，或点击 **「阅读原文」** 直达官网报名\r\n\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_01_zw7.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=1334rp2zSvHyyt9iQC%2Fldsg3FyI%3D)",{"publishTime":7,"createTime":109,"canonicalPath":114,"id":115,"title":116},"\u002Fstudy-material\u002Fstudy-10","study-10","InfiniTensor冬季训练营CPU并行编程第一课《CPU性能优化导论》",98,"\u002Fstudy-material\u002Fstudy-11","study-11","markdown",{"publishTime":7,"createTime":109,"canonicalPath":122,"id":123,"title":124},"\u002Fstudy-material\u002Fstudy-12","study-12","InfiniTensor 冬季训练营 Triton & 九齿第一课《Triton 练气术》",{"htmlContent":126,"headings":127},"\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_11_zw01.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=YIJ4V771ULsBRKLQqctuduXY3tc%3D\" alt=\"\">\u003C\u002Fp>\n\u003Cp>2026 年 1 月 13 日，2025 冬季 InfiniTensor 大模型与人工智能系统训练营训练系统系列第一课正式开讲，课程主题为《训练基础介绍》。\u003C\u002Fp>\n\u003Cp>本节课程主要介绍一些训练相关的基础知识，同时介绍一下团队正在开发的自研训练框架 InfiniTrain。\u003C\u002Fp>\n\u003Ch3 id=\"神经网络基础概念\">\u003Ca class=\"anchor\" href=\"#神经网络基础概念\" aria-label=\"跳转到标题：神经网络基础概念\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>神经网络基础概念\u003C\u002Fh3>\u003Cp>神经网络的基本组成结构\u003C\u002Fp>\n\u003Cul>\n\u003Cli>输入层 \u002F 隐藏层 \u002F 输出层\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_11_zw02.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=ZmYszhJUzPfNt4uBBZ%2Bysni2FhI%3D\" alt=\"\">\u003C\u002Fp>\n\u003Ch3 id=\"优化算法与损失函数\">\u003Ca class=\"anchor\" href=\"#优化算法与损失函数\" aria-label=\"跳转到标题：优化算法与损失函数\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>优化算法与损失函数\u003C\u002Fh3>\u003Ch4 id=\"1-梯度下降法\">\u003Ca class=\"anchor\" href=\"#1-梯度下降法\" aria-label=\"跳转到标题：1.梯度下降法\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>1.梯度下降法\u003C\u002Fh4>\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_11_zw03.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=u6swUoVx4fEzHj%2F%2FSj3XcBnRloo%3D\" alt=\"\">\u003C\u002Fp>\n\u003Ch4 id=\"2-损失函数类型\">\u003Ca class=\"anchor\" href=\"#2-损失函数类型\" aria-label=\"跳转到标题：2.损失函数类型\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>2.损失函数类型\u003C\u002Fh4>\u003Cp>交叉熵损失（分类任务）\u003Cbr>均方误差损失（回归任务）\u003C\u002Fp>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_11_zw04.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=wXAxMmLj%2B8%2F83StgNmmm4jzLJe0%3D\" alt=\"\">\u003C\u002Fp>\n\u003Ch3 id=\"激活函数详解\">\u003Ca class=\"anchor\" href=\"#激活函数详解\" aria-label=\"跳转到标题：激活函数详解\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>激活函数详解\u003C\u002Fh3>\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_11_zw05.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=xmitMfGBaBeQeWOv02qbYcWmzYA%3D\" alt=\"\">\u003C\u002Fp>\n\u003Ch3 id=\"反向传播的教学基础\">\u003Ca class=\"anchor\" href=\"#反向传播的教学基础\" aria-label=\"跳转到标题：反向传播的教学基础\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>反向传播的教学基础\u003C\u002Fh3>\u003Cul>\n\u003Cli>链式法则数学原理\u003C\u002Fli>\n\u003Cli>反向传播具体流程\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_11_zw06.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=VvHXK2QHwYUZ%2Ft%2FWI3mRTmvtRmU%3D\" alt=\"\">\u003C\u002Fp>\n\u003Ch3 id=\"pytorch-核心机制\">\u003Ca class=\"anchor\" href=\"#pytorch-核心机制\" aria-label=\"跳转到标题：PyTorch 核心机制\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>PyTorch 核心机制\u003C\u002Fh3>\u003Ch4 id=\"1-autograd-自动微分机制\">\u003Ca class=\"anchor\" href=\"#1-autograd-自动微分机制\" aria-label=\"跳转到标题：1.Autograd 自动微分机制\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>1.Autograd 自动微分机制\u003C\u002Fh4>\u003Ch4 id=\"2-构建-function-计算图\">\u003Ca class=\"anchor\" href=\"#2-构建-function-计算图\" aria-label=\"跳转到标题：2.构建 Function 计算图\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>2.构建 Function 计算图\u003C\u002Fh4>\u003Ch4 id=\"3-pytorch-梯度计算规则\">\u003Ca class=\"anchor\" href=\"#3-pytorch-梯度计算规则\" aria-label=\"跳转到标题：3.PyTorch 梯度计算规则\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>3.PyTorch 梯度计算规则\u003C\u002Fh4>\u003Ch4 id=\"4-requires-grad-参数作用\">\u003Ca class=\"anchor\" href=\"#4-requires-grad-参数作用\" aria-label=\"跳转到标题：4.requires_grad 参数作用\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>4.requires_grad 参数作用\u003C\u002Fh4>\u003Ch3 id=\"混合精度训练\">\u003Ca class=\"anchor\" href=\"#混合精度训练\" aria-label=\"跳转到标题：混合精度训练\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>混合精度训练\u003C\u002Fh3>\u003Cul>\n\u003Cli>torch.amp\u003C\u002Fli>\n\u003Cli>Nvidia APEX\u003C\u002Fli>\n\u003Cli>分层混合精度策略\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_11_zw07.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=iFV%2FPRx2Q7DJZp7Sdf3%2BYcWW0mo%3D\" alt=\"\">\u003C\u002Fp>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_11_zw08.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=vJtJEQ2xu%2FL42ajqaGVcuvm80Qo%3D\" alt=\"\">\u003C\u002Fp>\n\u003Ch3 id=\"分布式训练策略\">\u003Ca class=\"anchor\" href=\"#分布式训练策略\" aria-label=\"跳转到标题：分布式训练策略\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>分布式训练策略\u003C\u002Fh3>\u003Ch4 id=\"1-并行训练基础\">\u003Ca class=\"anchor\" href=\"#1-并行训练基础\" aria-label=\"跳转到标题：1.并行训练基础\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>1.并行训练基础\u003C\u002Fh4>\u003Cul>\n\u003Cli>数据并行原理\u003C\u002Fli>\n\u003Cli>模型并行策略\u003C\u002Fli>\n\u003Cli>混合并行方法\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch4 id=\"2-高级并行技术\">\u003Ca class=\"anchor\" href=\"#2-高级并行技术\" aria-label=\"跳转到标题：2.高级并行技术\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>2.高级并行技术\u003C\u002Fh4>\u003Cul>\n\u003Cli>张量并行实现\u003C\u002Fli>\n\u003Cli>序列并行优化\u003C\u002Fli>\n\u003Cli>流水线并行设计\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch3 id=\"显存优化技巧\">\u003Ca class=\"anchor\" href=\"#显存优化技巧\" aria-label=\"跳转到标题：显存优化技巧\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>显存优化技巧\u003C\u002Fh3>\u003Ch4 id=\"1-大模型训练挑战\">\u003Ca class=\"anchor\" href=\"#1-大模型训练挑战\" aria-label=\"跳转到标题：1.大模型训练挑战\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>1.大模型训练挑战\u003C\u002Fh4>\u003Cul>\n\u003Cli>显存占用分析\u003C\u002Fli>\n\u003Cli>优化策略概述\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch4 id=\"2-具体优化技术\">\u003Ca class=\"anchor\" href=\"#2-具体优化技术\" aria-label=\"跳转到标题：2.具体优化技术\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>2.具体优化技术\u003C\u002Fh4>\u003Cul>\n\u003Cli>ZeRO (Zero Redundancy Optimizer)\u003C\u002Fli>\n\u003Cli>梯度检查点机制\u003C\u002Fli>\n\u003Cli>显存高效训练方法\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch3 id=\"强化学习与-deepseek-入门\">\u003Ca class=\"anchor\" href=\"#强化学习与-deepseek-入门\" aria-label=\"跳转到标题：强化学习与 DeepSeek 入门\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>强化学习与 DeepSeek 入门\u003C\u002Fh3>\u003Ch4 id=\"1-强化学习基础\">\u003Ca class=\"anchor\" href=\"#1-强化学习基础\" aria-label=\"跳转到标题：1.强化学习基础\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>1.强化学习基础\u003C\u002Fh4>\u003Cul>\n\u003Cli>基本概念与原理\u003C\u002Fli>\n\u003Cli>常用算法介绍\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch4 id=\"2-deepseek\">\u003Ca class=\"anchor\" href=\"#2-deepseek\" aria-label=\"跳转到标题：2.DeepSeek\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>2.DeepSeek\u003C\u002Fh4>\u003Cul>\n\u003Cli>架构特点与应用\u003C\u002Fli>\n\u003Cli>实践入门指导\u003C\u002Fli>\n\u003Cli>实践环节与作业要求\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch3 id=\"自研框架-infinitrain\">\u003Ca class=\"anchor\" href=\"#自研框架-infinitrain\" aria-label=\"跳转到标题：自研框架 InfiniTrain\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>自研框架 InfiniTrain\u003C\u002Fh3>\u003Cul>\n\u003Cli>框架架构介绍\u003C\u002Fli>\n\u003Cli>Kernel 层（CUDA 内核）\u003C\u002Fli>\n\u003Cli>Autograd 层（自动微分）\u003C\u002Fli>\n\u003Cli>NN层（模型构建）\u003C\u002Fli>\n\u003Cli>示例模型实现\u003C\u002Fli>\n\u003Cli>GPT-2 训练示例\u003C\u002Fli>\n\u003Cli>LLaMA-3 训练示例\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch3 id=\"总结与展望\">\u003Ca class=\"anchor\" href=\"#总结与展望\" aria-label=\"跳转到标题：总结与展望\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>总结与展望\u003C\u002Fh3>\u003Cp>本节课深入理解前向传播、反向传播与梯度优化等训练核心基础。后续将系统学习分布式策略与自研框架实践，全面提升工程能力。\u003C\u002Fp>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_01_zw6.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=Jvvk9S3asMdwJQP5eMwtAGsqW0k%3D\" alt=\"\">\u003C\u002Fp>\n\u003Ch3 id=\"学习贴士\">\u003Ca class=\"anchor\" href=\"#学习贴士\" aria-label=\"跳转到标题：💡 学习贴士\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>💡 学习贴士\u003C\u002Fh3>\u003Cp>更多细节及完整课程内容，请观看直播或查看课程回放。\u003C\u002Fp>\n\u003Cp>链接：\u003Ca href=\"https:\u002F\u002Fwww.infinitensor.com\u002Fcamp\u002Fwinter2025\">https:\u002F\u002Fwww.infinitensor.com\u002Fcamp\u002Fwinter2025\u003C\u002Fa>\u003C\u002Fp>\n\u003Ch3 id=\"训练营报名持续进行中\">\u003Ca class=\"anchor\" href=\"#训练营报名持续进行中\" aria-label=\"跳转到标题：📚 训练营报名持续进行中\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>📚 训练营报名持续进行中\u003C\u002Fh3>\u003Cp>扫码报名与加入群聊，或点击 \u003Cstrong>「阅读原文」\u003C\u002Fstrong> 直达官网报名\u003C\u002Fp>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_01_zw7.jpg?Expires=1792752825&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=1334rp2zSvHyyt9iQC%2Fldsg3FyI%3D\" alt=\"\">\u003C\u002Fp>\n",[128,130,132,136,139,141,143,146,149,152,155,158,160,162,165,168,170,173,176,179,182,185,188,190,193],{"id":129,"text":129,"level":9},"神经网络基础概念",{"id":131,"text":131,"level":9},"优化算法与损失函数",{"id":133,"text":134,"level":135},"1-梯度下降法","1.梯度下降法",4,{"id":137,"text":138,"level":135},"2-损失函数类型","2.损失函数类型",{"id":140,"text":140,"level":9},"激活函数详解",{"id":142,"text":142,"level":9},"反向传播的教学基础",{"id":144,"text":145,"level":9},"pytorch-核心机制","PyTorch 核心机制",{"id":147,"text":148,"level":135},"1-autograd-自动微分机制","1.Autograd 自动微分机制",{"id":150,"text":151,"level":135},"2-构建-function-计算图","2.构建 Function 计算图",{"id":153,"text":154,"level":135},"3-pytorch-梯度计算规则","3.PyTorch 梯度计算规则",{"id":156,"text":157,"level":135},"4-requires-grad-参数作用","4.requires_grad 参数作用",{"id":159,"text":159,"level":9},"混合精度训练",{"id":161,"text":161,"level":9},"分布式训练策略",{"id":163,"text":164,"level":135},"1-并行训练基础","1.并行训练基础",{"id":166,"text":167,"level":135},"2-高级并行技术","2.高级并行技术",{"id":169,"text":169,"level":9},"显存优化技巧",{"id":171,"text":172,"level":135},"1-大模型训练挑战","1.大模型训练挑战",{"id":174,"text":175,"level":135},"2-具体优化技术","2.具体优化技术",{"id":177,"text":178,"level":9},"强化学习与-deepseek-入门","强化学习与 DeepSeek 入门",{"id":180,"text":181,"level":135},"1-强化学习基础","1.强化学习基础",{"id":183,"text":184,"level":135},"2-deepseek","2.DeepSeek",{"id":186,"text":187,"level":9},"自研框架-infinitrain","自研框架 InfiniTrain",{"id":189,"text":189,"level":9},"总结与展望",{"id":191,"text":192,"level":9},"学习贴士","💡 学习贴士",{"id":194,"text":195,"level":9},"训练营报名持续进行中","📚 训练营报名持续进行中"]