# deeplearning

**Repository Path**: techwolf/deeplearning

## Basic Information

- **Project Name**: deeplearning
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2025-12-16
- **Last Updated**: 2026-01-07

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# 人工智能与深度学习(Artificial Intelligence and Deep Learning)

讲解人工智能和深度学习的基础知识与方法实践，涵盖线性模型、多层感知器、卷积网络、循环神经网络和注意力机制等内容，适用于高年级本科生和研究生。

An introduction to the fundamentals and practical methods of artificial intelligence and deep learning, covering linear models, multi-layer perceptrons, convolutional networks, recurrent neural networks, and attention mechanisms. Suitable for senior undergraduates and graduate students.


# 课程结构(course structure)


## 预备知识(preliminary knowledge)

- 张量操作(tensor operations)
- 线性代数(linear algebra)
- 自动微分(automatic differentiation)

## 线性模型(linear models)

- 线性回归(linear regression)
- softmax回归(softmax regression)

## 多层感知器(multi-layer perceptrons)

- MLP(multi-layer perceptron)
- 权重衰减与正则(weight decay and regularization)
- 暂退法(dropout)
- 数值稳定性(numerical stability)

## 深度学习计算(deep learning computation)

- 模型结构(model architecture)
- 参数管理(parameter management)
- 自定义组件(custom components)
- 使用GPU(using GPU)

## 卷积网络(convolutional networks)

- 卷积基础(convolution basics)
- 卷积组件(convolution components)
- 现代网络架构(modern network architectures)

## 循环神经网络(recurrent neural networks)

- 语言模型(language models)
- 循环神经网络(recurrent neural networks)
- 长短期记忆模型(long short-term memory)
- 序列到序列模型(sequence-to-sequence models)


## 注意力机制(attention mechanisms)

- 注意力提示(attention cues)
- 注意力评分函数(attention scoring functions)
- 多头注意力(multihead attention)
- 自注意力和位置编码(self-attention and positional encoding)
- 变换器(transformers)
  

# 系统配置

## IDE

- Visual Studio Code
    - 安装：https://code.visualstudio.com/
    - 推荐插件
      - Python
      - Pylance
      - Jupyter
      - Markdown All in One
      - Markdown Preview Enhanced
      - Git Extension Pack
      - Optuna Dashboard
      - Quarto
- Jupyter Notebook / Jupyter Lab
  
## Python环境

- Miniconda
  - 安装：https://docs.anaconda.net.cn/miniconda/install/
  - 创建环境：`conda create -n Your-Env-Name python=3.10`
  - 激活环境：`conda activate Your-Env-Name`
  - 停用环境：`conda deactivate`
  - 删除环境：`conda remove -n Your-Env-Name --all`
  
- Python 3.8+


## 推荐安装包

### 基础计算

- numpy
- pandas
- scipy

### 可视化

- matplotlib
- seaborn

### 统计学与机器学习

- statmodels
- scikit-learn

### 深度学习

- pytorch
- torchvision

### 超参优化

- optuna
- imbalanced-learn

# Git配置
- 安装Git：https://git-scm.com/downloads
- 配置用户名和邮箱：
  ```
  git config --global user.name "Your Name"
  git config --global user.email "your.email@example.com"
  ```

## Git使用
- 克隆代码库：
  ```
  git clone <repository-url>
  ```
- 提交更改：
  ```
    git add .
    git commit -m "Your commit message"
    git push origin <branch-name>
    ```
- 拉取最新更改：
  ```
  git pull origin <branch-name>
  ```
- 创建分支：
  ```
  git checkout -b <new-branch-name>
  ```
- 切换分支：
  ```
  git checkout <branch-name>
  ```
- 合并分支：
  ```
  git checkout <target-branch>
  git merge <source-branch>
  ```
- 查看状态：
  ```
  git status
  ```
- 查看日志：
  ```
  git log
  ```
- 查看分支：
  ```
  git branch
  ```
- 删除分支：
  ```
  git branch -d <branch-name>
  ```
- 查看远程仓库：
  ```
  git remote -v
  ```
- 添加远程仓库：
  ```
  git remote add <remote-name> <repository-url>
  ```
  
## Git平台
- GitHub：https://github.com/
- GitLab：https://gitlab.com/
- GitCode：https://gitcode.com/
- Gitee：https://gitee.com/

# 课件获取

git clone https://gitcode.com/leelapis/deeplearning.git