From a98b1be31c3a1e595aa93c265e23a5d992d94ac1 Mon Sep 17 00:00:00 2001 From: Lanxiang Hu Date: Mon, 17 Mar 2025 00:02:22 +0000 Subject: [PATCH 1/4] add cse 234 w25 --- docs/机器学习系统/MLSYS.en.md | 28 +++++++++++++++++++++++++++ docs/机器学习系统/MLSYS.md | 36 +++++++++++++++++++++++++++++++++++ 2 files changed, 64 insertions(+) create mode 100644 docs/机器学习系统/MLSYS.en.md create mode 100644 docs/机器学习系统/MLSYS.md diff --git a/docs/机器学习系统/MLSYS.en.md b/docs/机器学习系统/MLSYS.en.md new file mode 100644 index 00000000..2674f8b2 --- /dev/null +++ b/docs/机器学习系统/MLSYS.en.md @@ -0,0 +1,28 @@ +# Data Systems for Machine Learning + +## Course Overview + +- University: University of California, San Diego +- Prerequisites: Fundamentals of Deep Learning and Computer Systems +- Programming Language: Python +- Course Difficulty: 🌟🌟🌟 +- Estimated Study Hours: 50 hours + +This course, offered in the Winter 2025 term by Professor Hao Zhang at the University of California, San Diego, focuses on machine learning systems, encompassing the latest research developments in large (langauge) models, machine learning compilation, and distributed systems. + +The curriculum is divided into three main sections: + +1. **Fundamentals**: Covers topics such as deep learning, automatic differentiation, and an overview of machine learning systems. + +2. **Machine Learning Systems and Optimization**: Includes subjects like dataflow graph systems, machine learning compilation, memory and graph optimizations, and distributed machine learning optimization. + +3. **Large (Language) Models**: Explores cutting-edge topics such as training of large language models (LLMs), data preparation, inference and serving, attention mechanism optimization, and retrieval-augmented generation (RAG). + +The course also features guest lectures from inventors of key technologies and industry leaders, providing students with direct interaction opportunities with experts. A foundation in deep learning and system programming is needed for this course. It offers extensive programming assignments and reading materials to help students deeply understand the design and optimization of machine learning systems. Self-learners should be aware that the course involves a significant amount of cutting-edge research, which may require additional time to consult related materials for a thorough understanding. + +## Course Resources + +- **Course Website**: [https://hao-ai-lab.github.io/cse234-w25/](https://hao-ai-lab.github.io/cse234-w25/) +- **Course Videos**: [https://podcast.ucsd.edu/watch/wi25/cse234_a00/1](https://podcast.ucsd.edu/watch/wi25/cse234_a00/1) +- **Course Notes**: [https://github.com/hao-ai-lab/cse234-w25/tree/main/assets/scribe_notes](https://github.com/hao-ai-lab/cse234-w25/tree/main/assets/scribe_notes) +- **Course Assignments**: [https://github.com/hao-ai-lab/cse234-w25-PA](https://github.com/hao-ai-lab/cse234-w25-PA) diff --git a/docs/机器学习系统/MLSYS.md b/docs/机器学习系统/MLSYS.md new file mode 100644 index 00000000..403d4f9d --- /dev/null +++ b/docs/机器学习系统/MLSYS.md @@ -0,0 +1,36 @@ +# Data Systems for Machine Learning + +## 课程简介 + +- 所属大学：加州大学圣迭戈分校 +- 先修要求：深度学习基础/计算机系统基础 +- 编程语言：Python +- 课程难度：🌟🌟🌟 +- 预计学时：50小时 + + + +这门课程由机器学习系统领域顶尖学者，来自加州大学圣迭戈分校的张昊教授于2025年冬季学期开设，聚焦于机器学习系统，涵盖大模型、机器学习编译和分布式系统等领域的最新研究进展。 + +课程内容分为三个部分： + +1. 基础知识：​包括深度学习、自动微分、机器学习系统概述等。 + +2. 机器学习系统与优化：​涵盖数据流图系统、机器学习编译、内存与图优化、分布式机器学习优化等主题。 + +3. 大（语言）模型：​探讨LLM的训练、数据准备、推理与服务、注意力机制优化、检索增强生成（RAG）等前沿话题。​ + +课程还邀请了多位关键技术的发明者和行业领军人物进行客座讲座，为学生提供与行业专家直接交流的机会。学习这门课程需要具备在深度学习和系统编程扎实的编程基础。​课程提供了丰富的编程作业和阅读材料，有助于学生深入理解机器学习系统的设计与优化。​自学者应注意，课程内容涉及大量前沿研究，可能需要额外时间查阅相关资料以加深理解。 + +## 课程资源 + +- 课程网站：https://hao-ai-lab.github.io/cse234-w25/ +- 课程视频：https://podcast.ucsd.edu/watch/wi25/cse234_a00/1 +- 课程笔记：https://github.com/hao-ai-lab/cse234-w25/tree/main/assets/scribe_notes +- 课程作业：https://github.com/hao-ai-lab/cse234-w25-PA \ No newline at end of file From 77a98b09ff921702e781c20e50991e336632c9d7 Mon Sep 17 00:00:00 2001 From: Lanxiang Hu Date: Mon, 17 Mar 2025 00:05:10 +0000 Subject: [PATCH 2/4] add cse 234 w25 --- docs/机器学习系统/MLSYS.en.md | 2 +- docs/机器学习系统/MLSYS.md | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/机器学习系统/MLSYS.en.md b/docs/机器学习系统/MLSYS.en.md index 2674f8b2..874a26fc 100644 --- a/docs/机器学习系统/MLSYS.en.md +++ b/docs/机器学习系统/MLSYS.en.md @@ -14,7 +14,7 @@ The curriculum is divided into three main sections: 1. **Fundamentals**: Covers topics such as deep learning, automatic differentiation, and an overview of machine learning systems. -2. **Machine Learning Systems and Optimization**: Includes subjects like dataflow graph systems, machine learning compilation, memory and graph optimizations, and distributed machine learning optimization. +2. **Machine Learning Systems and Optimization**: Includes subjects like machine learning compilation, memory and graph optimizations, and distributed machine learning optimization. 3. **Large (Language) Models**: Explores cutting-edge topics such as training of large language models (LLMs), data preparation, inference and serving, attention mechanism optimization, and retrieval-augmented generation (RAG). diff --git a/docs/机器学习系统/MLSYS.md b/docs/机器学习系统/MLSYS.md index 403d4f9d..261a5cb4 100644 --- a/docs/机器学习系统/MLSYS.md +++ b/docs/机器学习系统/MLSYS.md @@ -22,7 +22,7 @@ 1. 基础知识：​包括深度学习、自动微分、机器学习系统概述等。 -2. 机器学习系统与优化：​涵盖数据流图系统、机器学习编译、内存与图优化、分布式机器学习优化等主题。 +2. 机器学习系统与优化：机器学习编译、内存与图优化、分布式机器学习优化等主题。 3. 大（语言）模型：​探讨LLM的训练、数据准备、推理与服务、注意力机制优化、检索增强生成（RAG）等前沿话题。​ From bee196e1de22374a1ab8265657170df6dac70840 Mon Sep 17 00:00:00 2001 From: Lanxiang Hu Date: Mon, 17 Mar 2025 00:16:15 +0000 Subject: [PATCH 3/4] add cse 234 w25 --- docs/机器学习系统/MLSYS.en.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/机器学习系统/MLSYS.en.md b/docs/机器学习系统/MLSYS.en.md index 874a26fc..aeb51f38 100644 --- a/docs/机器学习系统/MLSYS.en.md +++ b/docs/机器学习系统/MLSYS.en.md @@ -8,7 +8,7 @@ - Course Difficulty: 🌟🌟🌟 - Estimated Study Hours: 50 hours -This course, offered in the Winter 2025 term by Professor Hao Zhang at the University of California, San Diego, focuses on machine learning systems, encompassing the latest research developments in large (langauge) models, machine learning compilation, and distributed systems. +This course, offered in the Winter 2025 quarter by Professor Hao Zhang at the University of California, San Diego, focuses on machine learning systems, encompassing the latest research developments in large (langauge) models, machine learning compilation, and distributed systems. The curriculum is divided into three main sections: From 28aad74ca997d1e977e99b56bc5db36bfcc646ff Mon Sep 17 00:00:00 2001 From: Lanxiang Hu Date: Mon, 17 Mar 2025 00:57:06 +0000 Subject: [PATCH 4/4] update cse234 course title --- docs/机器学习系统/MLSYS.en.md | 2 +- docs/机器学习系统/MLSYS.md | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/机器学习系统/MLSYS.en.md b/docs/机器学习系统/MLSYS.en.md index aeb51f38..c27a0b6e 100644 --- a/docs/机器学习系统/MLSYS.en.md +++ b/docs/机器学习系统/MLSYS.en.md @@ -1,4 +1,4 @@ -# Data Systems for Machine Learning +# Machine Learning Systems ## Course Overview diff --git a/docs/机器学习系统/MLSYS.md b/docs/机器学习系统/MLSYS.md index 261a5cb4..9bb5b6b9 100644 --- a/docs/机器学习系统/MLSYS.md +++ b/docs/机器学习系统/MLSYS.md @@ -1,4 +1,4 @@ -# Data Systems for Machine Learning +# Machine Learning Systems ## 课程简介