# dive_into_deep_learning **Repository Path**: zxmcode/dive_into_deep_learning ## Basic Information - **Project Name**: dive_into_deep_learning - **Description**: 动手学深度学习-李沐(.py文件版本) - **Primary Language**: Python - **License**: Not specified - **Default Branch**: main - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 1 - **Forks**: 0 - **Created**: 2021-11-24 - **Last Updated**: 2025-05-30 ## Categories & Tags **Categories**: Uncategorized **Tags**: None ## README # 动手学深度学习 李沐 dive-into-deep-learning 李沐老师的课程中源码都是用jupyter notebook写的;这里全部使用pycharm编辑器来编程,改写为py格式。 希望可以记录课程的学习过程,同时能帮助他人。 ### 课程相关资料 1. 课程的直播地址:http://courses.d2l.ai/zh-v2/ 2. 课程的课件地址:https://zh-v2.d2l.ai/ 3. 另一个可参考的笔记:https://tangshusen.me/Dive-into-DL-PyTorch ### 本笔记的目录 ##### ch01. 预备知识 1.1. [数据操作](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch01/01-ndarray.py) 1.2. [数据预处理](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch01/02-pandas.py) 1.3. [线性代数](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch01/03-linear-algebra.py) 1.4. [微分](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch01/04-calculus.py) 1.5. [自动求导](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch01/05-autograd.py) 1.6. 概率 ##### ch02. 线性神经网络 2.1. [线性回归](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch02/01-linear-regression.py) 2.2. [线性回归的从零开始实现](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch02/02-linear-regression-scratch.py) 2.3. 线性回归的简洁实现 2.4. softmax回归 2.5. [图像分类数据集](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/d2lutil/common.py) 2.6. [softmax回归的从零开始实现](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch02/03-softmax-linear-regression-scratch.py) 2.7. [softmax回归的简洁实现](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch02/04-softmax-linear-regression-concise.py) ##### ch03. 多层感知机 3.1. [多层感知机](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch03/01-mlp.py) 3.2. [多层感知机的从零开始实现](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch03/02-mlp-from-zero.py) 3.3. [多层感知机的简洁实现](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch03/03-mlp-simple.py) 3.4. [模型选择、欠拟合和过拟合](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch03/04-underfit-overfit.py) 3.5. [权重衰减](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch03/05-weight-decay-simple.py) 3.6. [Dropout](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch03/06-dropout-simple.py) 3.7. 正向传播、反向传播和计算图 3.8. 数值稳定性和模型初始化 3.9. 环境和分布偏移 3.10. 实战 Kaggle 比赛:预测房价 ##### ch04. 深度学习计算 4.1. [层和块](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch04/01-model-construction.py) 4.2. [参数管理](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch04/02-parameters.py) 4.3. 延后初始化 4.4. [自定义层](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch04/03-custom-layer.py) 4.5. [读写文件](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch04/04-read-write.py) 4.6. GPU ##### ch05. 卷积神经网络 5.1. 二维卷积层 5.2. 填充和步幅 5.3. [多输入多输出通道](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch05/03-channels.py) 5.4. [池化层](https://github.com/Miraclelucy/dive-into-deep-learning/blob/main/ch05/04-pooling.py) 5.5. 卷积神经网络(LeNet) 5.6. 深度卷积神经网络(AlexNet) 5.7. 使用重复元素的网络(VGG) 5.8. 网络中的网络(NiN) 5.9. 含并行连结的网络(GoogLeNet) 5.10. 批量归一化 5.11. 残差网络(ResNet) 5.12. 稠密连接网络(DenseNet) ##### ch06. 计算机视觉 6.1. 图像增广 6.2. 微调 6.3. 目标检测和边界框 6.4. 锚框 6.5. 多尺度目标检测 6.6. 目标检测数据集(皮卡丘) 6.7. 单发多框检测(SSD) 6.8. 区域卷积神经网络(R-CNN)系列 6.9. 语义分割和数据集 6.10. 转置卷积 6.11. 全卷积网络 6.12. 样式迁移 ##### ch07. 循环神经网络 7.1. 语言模型 7.2. 循环神经网络 7.3. 语言模型数据集(周杰伦专辑歌词) 7.4. 循环神经网络的从零开始实现 7.5. 循环神经网络的简洁实现 7.6. 通过时间反向传播 7.7. 门控循环单元(GRU) 7.8. 长短期记忆(LSTM) 7.9. 深度循环神经网络 7.10. 双向循环神经网络 7.11. 机器翻译与数据集 7.12. 编码器-解码器结构