{"id":14958786,"url":"https://github.com/zkywsg/daily-deeplearning","last_synced_at":"2025-05-15T08:05:18.577Z","repository":{"id":39364736,"uuid":"195398609","full_name":"zkywsg/Daily-DeepLearning","owner":"zkywsg","description":"🔥机器学习/深度学习/Python/大模型/多模态/LLM/deeplearning/Python/Algorithm interview/NLP Tutorial","archived":false,"fork":false,"pushed_at":"2025-03-02T18:26:27.000Z","size":52735,"stargazers_count":699,"open_issues_count":0,"forks_count":153,"subscribers_count":19,"default_branch":"master","last_synced_at":"2025-04-03T04:09:11.826Z","etag":null,"topics":["cv","deep-learning","leetcode","leetcode-python","leetcode-solutions","llm","machine-learning","nlp","python","pytorch","pytorch-nlp","pytorch-tutorial","pytorch-tutorials","tensorflow","tensorflow-examples","tensorflow-tutorials"],"latest_commit_sha":null,"homepage":"","language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/zkywsg.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2019-07-05T11:32:03.000Z","updated_at":"2025-04-02T12:35:55.000Z","dependencies_parsed_at":"2024-09-22T08:00:34.513Z","dependency_job_id":"06b64095-75ca-4ed0-97aa-f8972682f7f4","html_url":"https://github.com/zkywsg/Daily-DeepLearning","commit_stats":{"total_commits":150,"total_committers":2,"mean_commits":75.0,"dds":"0.026666666666666616","last_synced_commit":"9fa52be3c448f64ec574283474b04403a359fea2"},"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/zkywsg%2FDaily-DeepLearning","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/zkywsg%2FDaily-DeepLearning/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/zkywsg%2FDaily-DeepLearning/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/zkywsg%2FDaily-DeepLearning/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/zkywsg","download_url":"https://codeload.github.com/zkywsg/Daily-DeepLearning/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":248166865,"owners_count":21058481,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["cv","deep-learning","leetcode","leetcode-python","leetcode-solutions","llm","machine-learning","nlp","python","pytorch","pytorch-nlp","pytorch-tutorial","pytorch-tutorials","tensorflow","tensorflow-examples","tensorflow-tutorials"],"created_at":"2024-09-24T13:18:16.672Z","updated_at":"2025-05-15T08:05:18.566Z","avatar_url":"https://github.com/zkywsg.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"#  **Daily-DeepLearning** \n\n欢迎来到 **Daily-DearnLearning**，涵盖[计算机基础课程](07-BaseClass/)、[Python快速入门](01-Python/)、[数据科学包的使用](05-Machine-Learning-Code/数据分析工具/)、[机器学习](02-Machine-Learning/)、[深度学习](03-Deep-Learning/)、[自然语言处理](04-NLP/)、[LLM](08-LLM/)等。\n\n## 2017年：Attention is All you need\n\n**Background**\n\n在《Attention is all you need》之前，因为李飞飞教授推动的ImageNet数据集、GPU算力的提升，像CNN刚刚开始流行起来，如用Tensoflow或者Theano写一个手写数字识别。后来开始有人在NLP领域，用word2vec和LSTM的组合，在很多领域里做到SOTA的效果。后来就是2017年，由Google团队提出的这篇里程碑式的论文。\n\n[核心解析](08-LLM/Attentionisallyouneed/核心解析.md) | [论文链接](08-LLM/Attentionisallyouneed/attentionisallyouneed.pdf)  | [简单例子](08-LLM/Attentionisallyouneed/example.md) | [自注意力机制](08-LLM/Attentionisallyouneed/selfattention.md) | [多头注意力](08-LLM/Attentionisallyouneed/multihead.md) | [位置编码](08-LLM/Attentionisallyouneed/positionalencoding.md) | [Harvard NLP PyTorch实现Transformer](https://nlp.seas.harvard.edu/2018/04/03/attention.html) | [Transformer复现](08-LLM/Attentionisallyouneed/Transformer_code.md)\n\n## 2018年 BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding\n\n**Background**\n\nBert比较特殊的地方在于采用了双向上下文建模，通过掩码语言模型（Masked language Model），同时利用左右两侧上下文，解决传统模型中的单向性问题。还有很重要的一点，从Bert看来是，形成了“预训练+微调”的新范式，统一了多种NLP任务的框架，仅需在预训练模型基础上添加简单任务头即可适配下游任务。当时在11项NLP任务上刷新SOTA，开启了大规模预训练模型（Pre-trained Language Model, PLM）时代。\n\n[Bert解析](08-LLM/Bert/核心解析.md) | [论文链接](https://arxiv.org/abs/1810.04805) | [Bert复现](08-LLM/Bert/Bert_code.md)\n\n## 2018年 GPT1：Generative Pre-trained Transformer\n\n**Background**\n\n在NLP任务依赖定制化模型、传统单向语言模型（如LSTM）难以建模长距离上下文的背景下，GPT-1首次将Transformer解码器架构与无监督预训练结合，提出“生成式预训练+微调”范式。通过自回归预训练（预测下一个词）学习通用文本表示，仅需简单微调即可适配分类、推理等任务，在12项NLP任务中9项达到SOTA，验证了大模型规模化训练的潜力，为后续GPT系列奠定了基础。\n\n## 2018年  ELMo：Embeddings from Language Models\n\n**Background**\n\nELMo这个工作主要还是对词向量的改进，从静态的词向量转变成动态词向量，从而提升各项NLP任务上的性能。虽然和GPT、BERT在同一年的工作，但其实应该放在这两项工作前面的，从马后炮的角度来说，主要用的还是双向LSTM，相较于Transformer这样支持并行计算的架构，再配合上MLM来捕捉双向上下文。\n\n[ELMo解析](08-LLM/ELMo/核心解析.md) | [论文链接](https://arxiv.org/abs/1802.05365)\n\nMore....\n\n---\n\n### 🖥️ **计算机基础课程**  \n**数据结构**  \n\n[基本概念和算法评价](07-BaseClass/Ds/01基本概念和算法评价.md) | [线性表](07-BaseClass/Ds/02线性表.md) | [栈和队列](07-BaseClass/Ds/03栈和队列.md) | [树和二叉树](07-BaseClass/Ds/04树和二叉树.md) | [图](07-BaseClass/Ds/05图.md) | [查找](07-BaseClass/Ds/06查找.md) | [排序](07-BaseClass/Ds/07排序.md)  \n\n**操作系统**  \n\n[操作系统的基本概念](07-BaseClass/Os/01操作系统的基本概念.md) | [操作系统的发展和分类](07-BaseClass/Os/02操作系统的发展和分类.md) | [操作系统的运行环境](07-BaseClass/Os/03操作系统的运行环境.md) | [进程和线程](07-BaseClass/Os/04进程与线程.md) | [处理机调度](07-BaseClass/Os/05处理机调度.md) | [进程同步](07-BaseClass/Os/06进程同步.md) | [死锁](07-BaseClass/Os/07死锁.md) | [内容管理概念](07-BaseClass/Os/08内容管理概念.md) | [虚拟内存管理](07-BaseClass/Os/09虚拟内存管理.md) | [文件系统基础](07-BaseClass/Os/10文件系统基础.md)  \n\n**计算机网络**  \n\n[计算机网络概述](07-BaseClass/Cn/01计算机网络概述.md) | [计算机网络结构体系](07-BaseClass/Cn/02计算机网络结构体系.md) | [通信基础](07-BaseClass/Cn/03通信基础.md) | [奈氏准则和香农定理](07-BaseClass/Cn/04奈氏准则和香农定理.md) | [传输介质](07-BaseClass/Cn/05传输介质.md) | [物理层设备](07-BaseClass/Cn/06物理层设备.md) | [数据链路层的功能](07-BaseClass/Cn/07数据链路层的功能.md)  \n\n---\n\n### 🐍 **Python 快速入门**  \n[变量、字符串、数字和运算符](01-Python/Day01.md) |  [列表、元组](01-Python/Day02.md) |  [字典、集合](01-Python/Day03.md) |  [条件语句、循环](01-Python/Day04.md) | [函数的定义与调用](01-Python/Day05.md) | [迭代、生成器、迭代器](01-Python/Day06.md) | [高阶函数、装饰器](01-Python/Day07.md) | [面向对象编程](01-Python/Day08.md) | [类的高级特性](01-Python/Day09.md) | [错误处理与调试](01-Python/Day10.md) | [文件操作](01-Python/Day11.md) | [多线程与多进程](01-Python/Day12.md) | [日期时间、集合、结构体](01-Python/Day13.md) | [协程与异步编程](01-Python/Day14.md) | [综合实践](01-Python/Day15.md)  \n\n---\n\n### 📊 **数据科学包的使用**  \n**NumPy**  \n\n[创建 ndarray](05-Machine-Learning-Code/数据分析工具/Numpy/创建ndarray.md) | [数据类型和运算](05-Machine-Learning-Code/数据分析工具/Numpy/数据类型和运算.md) | [索引和切片](05-Machine-Learning-Code/数据分析工具/Numpy/索引和切片.md) | [矩阵操作](05-Machine-Learning-Code/数据分析工具/Numpy/矩阵操作.md)  \n\n**Pandas**  \n\n[加载数据](05-Machine-Learning-Code/数据分析工具/Pandas/1_Loading.ipynb) | [行列选择](05-Machine-Learning-Code/数据分析工具/Pandas/2_Select_row_and_columns.ipynb) | [索引操作](05-Machine-Learning-Code/数据分析工具/Pandas/3_Set_reset_use_indexes.ipynb) | [数据过滤](05-Machine-Learning-Code/数据分析工具/Pandas/4_Filtering.ipynb) | [更新行列](05-Machine-Learning-Code/数据分析工具/Pandas/5_update_rows_columns.ipynb) | [数据排序](05-Machine-Learning-Code/数据分析工具/Pandas/7_sort_data.ipynb) | [数据聚合](05-Machine-Learning-Code/数据分析工具/Pandas/8_Grouping_Aggregating.ipynb) | [数据清洗](05-Machine-Learning-Code/数据分析工具/Pandas/9_Cleaning_Data.ipynb) | [时间数据处理](05-Machine-Learning-Code/数据分析工具/Pandas/10_WorkingWithDatesAndTimeSertesData.ipynb)  \n\n**Matplotlib**  \n\n[直线图](05-Machine-Learning-Code/数据分析工具/Matplotlib/1_creating_and_customizing_plots.ipynb) | [柱状图](05-Machine-Learning-Code/数据分析工具/Matplotlib/2_Bar_charts.ipynb) | [饼状图](05-Machine-Learning-Code/数据分析工具/Matplotlib/3_Pie.ipynb) | [堆叠图](05-Machine-Learning-Code/数据分析工具/Matplotlib/4_stack.ipynb) | [填充图](05-Machine-Learning-Code/数据分析工具/Matplotlib/5_Line_Filling_Area.ipynb) | [直方图](05-Machine-Learning-Code/数据分析工具/Matplotlib/6_histograms.ipynb) | [散点图](05-Machine-Learning-Code/数据分析工具/Matplotlib/7_Scatter.ipynb) | [时序图](05-Machine-Learning-Code/数据分析工具/Matplotlib/8_Time_Series_Data.ipynb) | [子图](05-Machine-Learning-Code/数据分析工具/Matplotlib/10_subplot.ipynb)  \n\n---\n\n### 🤖 **机器学习理论与实战**  \n**理论**  \n\n[逻辑回归](02-Machine-Learning/逻辑回归.md) | [EM 算法](02-Machine-Learning/EM算法.md) | [集成学习](02-Machine-Learning/集成学习入门.md) | [随机森林与 GBDT](02-Machine-Learning/随机森林和GBDT.md) | [ID3/C4.5 算法](02-Machine-Learning/ID3和C4.5算法.md) | [K-means](02-Machine-Learning/K-means.md) | [K 最近邻](02-Machine-Learning/K最近邻.md) | [贝叶斯](02-Machine-Learning/贝叶斯.md) | [XGBoost/LightGBM](02-Machine-Learning/XgBoost和LightGBM.md) | [Gradient Boosting](02-Machine-Learning/Gradient_Boosting.md) | [Boosting Tree](https://mp.weixin.qq.com/s/Cdi0CcWDLgS6Kk7Kx71Vaw) | [回归树](https://mp.weixin.qq.com/s/XiTH-8FY5Aw-p_1Ifhx4oQ) | [XGBoost](02-Machine-Learning/XgBoost.md) | [GBDT 分类](02-Machine-Learning/GBDT分类.md) | [GBDT 回归](02-Machine-Learning/GBDT回归.md) | [LightGBM](02-Machine-Learning/LightGBM.md) | [CatBoost](02-Machine-Learning/CatBoost.md)  \n\n---\n\n### 🏊‍♀️ **深度学习理论与实战**  \n**理论**  \n\n[Word2Vec](03-Deep-Learning/Word2Vec.md) | [BatchNorm](03-Deep-Learning/BatchNorm.md) | [Dropout](03-Deep-Learning/Dropout.md) | [CNN](03-Deep-Learning/CNN.md) | [RNN](03-Deep-Learning/RNN.md) | [LSTM](03-Deep-Learning/LSTM.md) | [Attention](03-Deep-Learning/Attention.md) | [ELMo](03-Deep-Learning/ELMo.md) | [Transformer](03-Deep-Learning/Transformer.md) | [BERT](03-Deep-Learning/BERT.md) | [ALBERT](03-Deep-Learning/ALBERT.md) | [XLNet](03-Deep-Learning/XLNet.md)  \n\n**实战**  \n\n- **TensorFlow**  \n  \n  [Hello World](06-Deep-Learning-Code/Tensorflow/Helloworld.md) | [线性回归](06-Deep-Learning-Code/Tensorflow/linear_regression.md) | [逻辑回归](06-Deep-Learning-Code/Tensorflow/logistic_regression.md) | [基本图像分类](06-Deep-Learning-Code/Tensorflow/基本图像分类.ipynb)  \n- **PyTorch**  \n  \n  [入门](06-Deep-Learning-Code/pytorch/gettingstart.md) | [自动求导](06-Deep-Learning-Code/pytorch/autograd.ipynb) | [神经网络](06-Deep-Learning-Code/pytorch/NeuralNetworks.ipynb)  \n\n---\n\n### 🀄 **NLP 相关**  \n[Word2Vec](03-Deep-Learning/Word2Vec.md) | [LSTM](03-Deep-Learning/LSTM.md) | [ELMo](03-Deep-Learning/ELMo.md) | [ALBERT](03-Deep-Learning/ALBERT.md) | [XLNet](03-Deep-Learning/XLNet.md)  \n\n---\n\n### 📫 **联系我们**  \n\n如果你有任何问题或建议，欢迎通过以下方式联系我们：  \n\n- **邮箱**：[lauzanhing@gmail.com](mailto:lauzanhing@gmail.com)  \n- **GitHub Issues**：[https://github.com/yourusername/Daily-DearnLearning/issues](https://github.com/yourusername/Daily-DearnLearning/issues)  \n\n---\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fzkywsg%2Fdaily-deeplearning","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fzkywsg%2Fdaily-deeplearning","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fzkywsg%2Fdaily-deeplearning/lists"}