Index

CMU 10-414/714: Deep Learning Systems¶

智能计算系统 - CS自学指南
学哪个待定, 先看看introduction TODO:
CMU 11-868: Large Language Model System - CS自学指南

prompt¶

我有一份英文课程 PPT。请根据 PPT 内容，翻译出一份完整的中文笔记，要求如下：

- 用 `markdown` 代码块输出
- 公式使用 LaTeX 语法（用 `$` 包裹, `$`符号和公式之间不要有空格）
- 笔记需要涵盖 PPT 的**所有**内容，包括概念、定理、例子、图表信息等, 如果需要插入截图, 请标注
- 每行末尾不需要加上`。`
- 结构清晰，适当使用标题（`#`）、代码块、内嵌latex公式等 Markdown 语法, 其中标题为# Lec6 xxx, ## 6.1, ## 6.2...

Schedule¶

Homework Details¶

Homework	Released	Due	Colab Link	GitHub Repo
Homework 0	8/29/24	9/12/24	hw0 colab	hw0 github
Homework 1	9/10/24	9/24/24	hw1 colab	hw1 github
Homework 2	9/24/24	10/10/24	hw2 colab	hw2 github
Homework 3	10/10/24	10/31/24	hw3 colab	hw3 github
Homework 4	10/31/24	11/16/24	hw4 colab	hw4 github
Homework 4 Extra (714 only)	10/31/24	11/16/24	hw4 extra colab	hw4 extra github

Lecture&Homework Schedule¶

Date (CMU)	Lecture	Instructor	Slides	Video (2022 version)	Homework
8/27	1 - Introduction / Logistics	Kolter	pdf	YouTube / Bilibili
8/29	2 - ML Refresher / Softmax Regression	Kolter	pdf	YouTube / Bilibili	Homework 0 Released
9/3	3 - Manual Neural Networks / Backprop	Kolter	pdf	YouTube (pt 1) YouTube (pt 2) / Bilibili (pt 1) Bilibili (pt 2)
9/5	4 - Automatic Differentiation	Chen	pdf	YouTube / Bilibili
9/10	5 - Automatic Differentiation Implementation	Chen	ipynb	YouTube / Bilibili	Homework 1 Released
9/12	6 - Optimization	Kolter	pdf	YouTube / Bilibili	Homework 0 Due
9/17	7 - Neural Network Library Abstractions	Chen	pdf	YouTube / Bilibili
9/19	9 - Normalization, Dropout, + Implementation	Kolter	pdf	YouTube / Bilibili
9/24	8 - NN Library Implementation	Chen	ipynb	YouTube / Bilibili	Homework 1 Due, Homework 2 Released
9/26	10 - Convolutional Networks	Kolter	pdf	YouTube / Bilibili
10/1	11 - Hardware Acceleration for Linear Algebra	Chen	pdf	YouTube / Bilibili
10/3	12 - Hardware Acceleration + GPUs	Chen	pdf	YouTube / Bilibili
10/8	13 - Hardware Acceleration Implementation	Chen	ipynb	YouTube / Bilibili
10/10	14 - Convoluations Network Implementation	Kolter	ipynb	YouTube / Bilibili	Homework 2 Due, Homework 3 Released
10/15	No class - Fall Break
10/17	No class - Fall Break
10/22	15 - Sequence Modeling + RNNs	Kolter	pdf	YouTube / Bilibili
10/24	16 - Sequence Modeling Implementation	Kolter	ipynb	YouTube / Bilibili
10/29	17 - Transformers and Autoregressive Models	Kolter	pdf	Youtube / Bilibili
10/31	18 - Transformers Implementation	Kolter	ipynb	Youtube / Bilibili	Homework 3 Due, Homework 4 Released, Homework 4 Extra Released
11/5	No class - Democracy Day
11/7	19 - Training Large Models	Chen	pdf	YouTube / Bilibili
11/12	20 - Generative Models	Chen	pdf	YouTube / Bilibili
11/14	21 - Generative Models Implementation	Chen	ipynb	YouTube / Bilibili
11/16					Homework 4 Due, Homework 4 Extra Due
11/19	22 - Customize Pretrained Models	Chen	pdf
11/21	23 - Model Deployment	Chen	pdf	Youtube / Bilibili
11/26	24 - Machine Learning Compilation and Deployment Implementation	Chen	ipynb	Youtube / Bilibili
11/28	No class - Thanksgiving
12/3	25 - Future Directions / Q&A	Both
12/5	26 - Student project presentations	Students

Slides下载¶

AI真是太好用了XD

#!/bin/bash

# 创建一个目录来存放下载的幻灯片
OUTPUT_DIR="dlsys_course_slides"
mkdir -p "$OUTPUT_DIR"
echo "将在 '$OUTPUT_DIR/' 目录中保存所有文件"
echo "-----------------------------------------"

# 定义一个辅助函数来处理单个文件的下载，避免代码重复
download_file() {
    local filename="$1"
    local url="$2"

    echo "正在下载: $filename"
    # 使用 wget 命令
    # -O 指定输出文件名 (包括目录)
    # -q 使其静默运行，--show-progress 显示一个简洁的进度条
    wget -q --show-progress -O "$OUTPUT_DIR/$filename" "$url"

    if [ $? -eq 0 ]; then
        echo "下载完成."
    else
        echo "下载失败: $filename"
    fi
    echo "" # 添加一个空行以提高可读性
}

# --- 所有 PDF 链接和对应的文件名都硬编码在此 ---

download_file "1 - Introduction - Logistics.pdf" "https://dlsyscourse.org/slides/intro.pdf"
download_file "2 - ML Refresher - Softmax Regression.pdf" "https://dlsyscourse.org/slides/2-softmax_regression.pdf"
download_file "3 - Manual Neural Networks - Backprop.pdf" "https://dlsyscourse.org/slides/manual_neural_nets.pdf"
download_file "4 - Automatic Differentiation.pdf" "https://dlsyscourse.org/slides/4-automatic-differentiation.pdf"
download_file "6 - Optimization.pdf" "https://dlsyscourse.org/slides/fc_init_opt.pdf"
download_file "7 - Neural Network Library Abstractions.pdf" "https://dlsyscourse.org/slides/7-nn-framework.pdf"
download_file "9 - Normalization Dropout Implementation.pdf" "https://dlsyscourse.org/slides/norm_reg.pdf"
download_file "10 - Convolutional Networks.pdf" "https://dlsyscourse.org/slides/conv_nets.pdf"
download_file "11 - Hardware Acceleration for Linear Algebra.pdf" "https://dlsyscourse.org/slides/11-hardware-acceleration.pdf"
download_file "12 - Hardware Acceleration + GPUs.pdf" "https://dlsyscourse.org/slides/12-gpu-acceleration.pdf"
download_file "15 - Sequence Modeling + RNNs.pdf" "https://dlsyscourse.org/slides/rnns.pdf"
download_file "17 - Transformers and Autoregressive Models.pdf" "https://dlsyscourse.org/slides/transformers.pdf"
download_file "19 - Training Large Models.pdf" "https://dlsyscourse.org/slides/15-training-large-models.pdf"
download_file "20 - Generative Models.pdf" "https://dlsyscourse.org/slides/16-generative-models.pdf"
download_file "22 - Customize Pretrained Models.pdf" "https://dlsyscourse.org/slides/22-augment-pretrained-models.pdf"
download_file "23 - Model Deployment.pdf" "https://dlsyscourse.org/slides/23-model-deployment.pdf"

# --- 下载列表结束 ---

echo "-----------------------------------------"
echo "所有预设的 PDF 均已下载完毕"