[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-brianspiering--awesome-dl4nlp":3,"tool-brianspiering--awesome-dl4nlp":64},[4,17,25,39,48,56],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":16},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",138956,2,"2026-04-05T11:33:21",[13,14,15],"开发框架","Agent","语言模型","ready",{"id":18,"name":19,"github_repo":20,"description_zh":21,"stars":22,"difficulty_score":10,"last_commit_at":23,"category_tags":24,"status":16},3704,"NextChat","ChatGPTNextWeb\u002FNextChat","NextChat 是一款轻量且极速的 AI 助手，旨在为用户提供流畅、跨平台的大模型交互体验。它完美解决了用户在多设备间切换时难以保持对话连续性，以及面对众多 AI 模型不知如何统一管理的痛点。无论是日常办公、学习辅助还是创意激发，NextChat 都能让用户随时随地通过网页、iOS、Android、Windows、MacOS 或 Linux 端无缝接入智能服务。\n\n这款工具非常适合普通用户、学生、职场人士以及需要私有化部署的企业团队使用。对于开发者而言，它也提供了便捷的自托管方案，支持一键部署到 Vercel 或 Zeabur 等平台。\n\nNextChat 的核心亮点在于其广泛的模型兼容性，原生支持 Claude、DeepSeek、GPT-4 及 Gemini Pro 等主流大模型，让用户在一个界面即可自由切换不同 AI 能力。此外，它还率先支持 MCP（Model Context Protocol）协议，增强了上下文处理能力。针对企业用户，NextChat 提供专业版解决方案，具备品牌定制、细粒度权限控制、内部知识库整合及安全审计等功能，满足公司对数据隐私和个性化管理的高标准要求。",87618,"2026-04-05T07:20:52",[13,15],{"id":26,"name":27,"github_repo":28,"description_zh":29,"stars":30,"difficulty_score":10,"last_commit_at":31,"category_tags":32,"status":16},2268,"ML-For-Beginners","microsoft\u002FML-For-Beginners","ML-For-Beginners 是由微软推出的一套系统化机器学习入门课程，旨在帮助零基础用户轻松掌握经典机器学习知识。这套课程将学习路径规划为 12 周，包含 26 节精炼课程和 52 道配套测验，内容涵盖从基础概念到实际应用的完整流程，有效解决了初学者面对庞大知识体系时无从下手、缺乏结构化指导的痛点。\n\n无论是希望转型的开发者、需要补充算法背景的研究人员，还是对人工智能充满好奇的普通爱好者，都能从中受益。课程不仅提供了清晰的理论讲解，还强调动手实践，让用户在循序渐进中建立扎实的技能基础。其独特的亮点在于强大的多语言支持，通过自动化机制提供了包括简体中文在内的 50 多种语言版本，极大地降低了全球不同背景用户的学习门槛。此外，项目采用开源协作模式，社区活跃且内容持续更新，确保学习者能获取前沿且准确的技术资讯。如果你正寻找一条清晰、友好且专业的机器学习入门之路，ML-For-Beginners 将是理想的起点。",84991,"2026-04-05T10:45:23",[33,34,35,36,14,37,15,13,38],"图像","数据工具","视频","插件","其他","音频",{"id":40,"name":41,"github_repo":42,"description_zh":43,"stars":44,"difficulty_score":45,"last_commit_at":46,"category_tags":47,"status":16},3128,"ragflow","infiniflow\u002Fragflow","RAGFlow 是一款领先的开源检索增强生成（RAG）引擎，旨在为大语言模型构建更精准、可靠的上下文层。它巧妙地将前沿的 RAG 技术与智能体（Agent）能力相结合，不仅支持从各类文档中高效提取知识，还能让模型基于这些知识进行逻辑推理和任务执行。\n\n在大模型应用中，幻觉问题和知识滞后是常见痛点。RAGFlow 通过深度解析复杂文档结构（如表格、图表及混合排版），显著提升了信息检索的准确度，从而有效减少模型“胡编乱造”的现象，确保回答既有据可依又具备时效性。其内置的智能体机制更进一步，使系统不仅能回答问题，还能自主规划步骤解决复杂问题。\n\n这款工具特别适合开发者、企业技术团队以及 AI 研究人员使用。无论是希望快速搭建私有知识库问答系统，还是致力于探索大模型在垂直领域落地的创新者，都能从中受益。RAGFlow 提供了可视化的工作流编排界面和灵活的 API 接口，既降低了非算法背景用户的上手门槛，也满足了专业开发者对系统深度定制的需求。作为基于 Apache 2.0 协议开源的项目，它正成为连接通用大模型与行业专有知识之间的重要桥梁。",77062,3,"2026-04-04T04:44:48",[14,33,13,15,37],{"id":49,"name":50,"github_repo":51,"description_zh":52,"stars":53,"difficulty_score":45,"last_commit_at":54,"category_tags":55,"status":16},519,"PaddleOCR","PaddlePaddle\u002FPaddleOCR","PaddleOCR 是一款基于百度飞桨框架开发的高性能开源光学字符识别工具包。它的核心能力是将图片、PDF 等文档中的文字提取出来，转换成计算机可读取的结构化数据，让机器真正“看懂”图文内容。\n\n面对海量纸质或电子文档，PaddleOCR 解决了人工录入效率低、数字化成本高的问题。尤其在人工智能领域，它扮演着连接图像与大型语言模型（LLM）的桥梁角色，能将视觉信息直接转化为文本输入，助力智能问答、文档分析等应用场景落地。\n\nPaddleOCR 适合开发者、算法研究人员以及有文档自动化需求的普通用户。其技术优势十分明显：不仅支持全球 100 多种语言的识别，还能在 Windows、Linux、macOS 等多个系统上运行，并灵活适配 CPU、GPU、NPU 等各类硬件。作为一个轻量级且社区活跃的开源项目，PaddleOCR 既能满足快速集成的需求，也能支撑前沿的视觉语言研究，是处理文字识别任务的理想选择。",74913,"2026-04-05T10:44:17",[15,33,13,37],{"id":57,"name":58,"github_repo":59,"description_zh":60,"stars":61,"difficulty_score":45,"last_commit_at":62,"category_tags":63,"status":16},2181,"OpenHands","OpenHands\u002FOpenHands","OpenHands 是一个专注于 AI 驱动开发的开源平台，旨在让智能体（Agent）像人类开发者一样理解、编写和调试代码。它解决了传统编程中重复性劳动多、环境配置复杂以及人机协作效率低等痛点，通过自动化流程显著提升开发速度。\n\n无论是希望提升编码效率的软件工程师、探索智能体技术的研究人员，还是需要快速原型验证的技术团队，都能从中受益。OpenHands 提供了灵活多样的使用方式：既可以通过命令行（CLI）或本地图形界面在个人电脑上轻松上手，体验类似 Devin 的流畅交互；也能利用其强大的 Python SDK 自定义智能体逻辑，甚至在云端大规模部署上千个智能体并行工作。\n\n其核心技术亮点在于模块化的软件智能体 SDK，这不仅构成了平台的引擎，还支持高度可组合的开发模式。此外，OpenHands 在 SWE-bench 基准测试中取得了 77.6% 的优异成绩，证明了其解决真实世界软件工程问题的能力。平台还具备完善的企业级功能，支持与 Slack、Jira 等工具集成，并提供细粒度的权限管理，适合从个人开发者到大型企业的各类用户场景。",70612,"2026-04-05T11:12:22",[15,14,13,36],{"id":65,"github_repo":66,"name":67,"description_en":68,"description_zh":69,"ai_summary_zh":69,"readme_en":70,"readme_zh":71,"quickstart_zh":72,"use_case_zh":73,"hero_image_url":74,"owner_login":75,"owner_name":76,"owner_avatar_url":77,"owner_bio":78,"owner_company":79,"owner_location":80,"owner_email":81,"owner_twitter":79,"owner_website":79,"owner_url":82,"languages":79,"stars":83,"forks":84,"last_commit_at":85,"license":79,"difficulty_score":86,"env_os":87,"env_gpu":88,"env_ram":88,"env_deps":89,"category_tags":92,"github_topics":79,"view_count":10,"oss_zip_url":79,"oss_zip_packed_at":79,"status":16,"created_at":93,"updated_at":94,"faqs":95,"releases":96},2730,"brianspiering\u002Fawesome-dl4nlp","awesome-dl4nlp","A curated list of awesome Deep Learning (DL) for Natural Language Processing (NLP) resources","awesome-dl4nlp 是一个精心整理的深度学习与自然语言处理（NLP）资源清单，旨在为学习者和技术人员提供一站式的入门与进阶指南。面对该领域知识更新快、资料分散的痛点，它将全球顶尖的课程视频、经典书籍、实战教程、前沿论文、开源框架及数据集进行了系统化分类与汇总。\n\n无论是希望系统构建知识体系的初学者，还是急需查找特定模型实现或最新研究成果的开发者与科研人员，都能在这里快速定位高质量内容。其独特亮点在于收录了斯坦福 CS224N 等世界名校的完整课程资料，以及从基础词向量到复杂神经网络模型的多种代码实现，覆盖了从理论数学推导到 Python 实战落地的全链路需求。通过这份清单，用户可以高效地跳过信息筛选的繁琐过程，直接触达社区公认的最佳实践，从而更专注于技术探索与应用创新。","Awesome Deep Learning for Natural Language Processing (NLP) [![Awesome](https:\u002F\u002Fcdn.rawgit.com\u002Fsindresorhus\u002Fawesome\u002Fd7305f38d29fed78fa85652e3a63e154dd8e8829\u002Fmedia\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Fsindresorhus\u002Fawesome)\n====\n\nTable of Contents\n----\n\n- __[Courses](#courses)__\n- __[Books](#books)__\n- __[Tutorials](#tutorials)__\n- __[Talks \u002F Lectures](#talks)__\n- __[Frameworks \u002F Models](#frameworks)__\n- __[Papers](#papers)__\n- __[Blog Posts](#blog-posts)__\n- __[Datasets](#datasets)__\n- __[Word Embeddings \u002F Word Vectors](#word-embeddings)__\n- __[Contributing](#contributing)__\n\nCourses\n----\n1. NLP with Deep Learning \u002F CS224N from Stanford (Winter 2019)\n\t- [Course homepage](http:\u002F\u002Fweb.stanford.edu\u002Fclass\u002Fcs224n\u002Findex.html) A complete survey of the field with videos, lecture slides, and sample student projects.\n\t- [Course lectures](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=8rXD5-xhemo&list=PLoROMvodv4rOhcuXMZkNm7j3fVwBBY42z) Video playlist.\n\t- [Previous course notes](https:\u002F\u002Fgithub.com\u002Fstanfordnlp\u002Fcs224n-winter17-notes) Probably the best \"book\" on DL for NLP.\n\t- [Course code](https:\u002F\u002Fgithub.com\u002FDSKSD\u002FDeepNLP-models-Pytorch) Pytorch implementations of various Deep NLP models in cs-224n.\n1. Neural Networks for NLP from Carnegie Mellon University\n\t- [Course homepage](http:\u002F\u002Fphontron.com\u002Fclass\u002Fnn4nlp2017\u002F)\n\t- [Course lectures](https:\u002F\u002Fwww.youtube.com\u002Fuser\u002Fneubig\u002Fvideos)\n\t- [Course code](https:\u002F\u002Fgithub.com\u002Fneubig\u002Fnn4nlp2017-code\u002F)\n1. Deep Learning for Natural Language Processing from University of Oxford and DeepMind\n\t- [Course homepage](https:\u002F\u002Fwww.cs.ox.ac.uk\u002Fteaching\u002Fcourses\u002F2016-2017\u002Fdl\u002F)\n\t- [Course slides](https:\u002F\u002Fgithub.com\u002Foxford-cs-deepnlp-2017\u002Flectures)\n\t- [Course lectures](https:\u002F\u002Fwww.youtube.com\u002Fplaylist?list=PL613dYIGMXoZBtZhbyiBqb0QtgK6oJbpm)\n\nBooks\n-----\n1. [Deep Learning with Text: Natural Language Processing (Almost) from Scratch with Python and spaCy](https:\u002F\u002Fwww.amazon.com\u002FDeep-Learning-Text-Approach-Processing\u002Fdp\u002F1491984414) by Patrick Harrison and Matthew Honnibal \n1. [Neural Network Methods in Natural Language Processing](https:\u002F\u002Fwww.amazon.com\u002Fgp\u002Fproduct\u002F1627052984) by Yoav Goldberg and Graeme Hirst\n1. [Deep Learning in Natural Language Processing](http:\u002F\u002Fwww.springer.com\u002Fus\u002Fbook\u002F9789811052088) by Li Deng and Yang Liu\n1. [The Math Behind Artificial Intelligence: A Guide to AI Foundations [Full Book]](https:\u002F\u002Fwww.freecodecamp.org\u002Fnews\u002Fthe-math-behind-artificial-intelligence-book\u002F) - Free book teaching the math behind AI with plain English explanations and Python code examples\n1. [Natural Language Processing in Action](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fnatural-language-processing-in-action) by Hobson Lane, Cole Howard, and Hannes Hapke\n1. Deep Learning: Natural Language Processing in Python by The LazyProgrammer (Kindle only)\n\t1. [Word2Vec and Word Embeddings in Python and Theano](https:\u002F\u002Fwww.amazon.com\u002FDeep-Learning-Language-Processing-Embeddings-ebook\u002Fdp\u002FB01KQ0ZN0A)\n\t1. [From Word2Vec to GLoVe in Python and Theano](https:\u002F\u002Fwww.amazon.com\u002FDeep-Learning-Language-Processing-Word2Vec-ebook\u002Fdp\u002FB01KRBOO4Y\u002F)\n\t1. [Recursive Neural Networks: Recursive Neural (Tensor) Networks in Theano](https:\u002F\u002Fwww.amazon.com\u002FDeep-Learning-Language-Processing-Recursive-ebook\u002Fdp\u002FB01KS5AEXO)\n1. [Applied Natural Language Processing with Python](https:\u002F\u002Fwww.amazon.ca\u002FApplied-Natural-Language-Processing-Python\u002Fdp\u002F1484237323) by  Taweh Beysolow II\n1. [Deep Learning Cookbook](https:\u002F\u002Fwww.amazon.ca\u002FDeep-Learning-Cookbook-Practical-Recipes\u002Fdp\u002F149199584X) by Douwe Osinga\n1. [Deep Learning for Natural Language Processing: Creating Neural Networks with Python](https:\u002F\u002Fwww.amazon.ca\u002FDeep-Learning-Natural-Language-Processing\u002Fdp\u002F148423684X) by Palash Goyal, Sumit Pandey, Karan Jain\n1. [Machine Learning for Text](https:\u002F\u002Fwww.amazon.ca\u002FMachine-Learning-Text-Charu-Aggarwal\u002Fdp\u002F3319735306) by Charu C. Aggarwal\n1. [Natural Language Processing with TensorFlow](https:\u002F\u002Fwww.amazon.ca\u002FNatural-Language-Processing-TensorFlow-language-ebook\u002Fdp\u002FB077Q3VZFR) by Thushan Ganegedara\n1. [fastText Quick Start Guide: Get started with Facebook's library for text representation and classification](https:\u002F\u002Fwww.amazon.ca\u002FfastText-Quick-Start-Guide-representation\u002Fdp\u002F1789130999)\n1. [Hands-On Natural Language Processing with Python](https:\u002F\u002Fwww.amazon.ca\u002FHands-Natural-Language-Processing-Python\u002Fdp\u002F178913949X)\n1. [Natural Language Processing in Action, Seond Edition](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fnatural-language-processing-in-action-second-edition) by Hobson Lane and Maria Dyshel\n1. [Getting Started with Natural Language Processing in Action](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fgetting-started-with-natural-language-processing) by Ekaterina Kochmar\n2. [Deep Learning for Natural Language Processing in Action](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fdeep-learning-for-natural-language-processing) by Stephan Raaijmakers\n\nTutorials\n-----\n\n1. [Text classification guide from Google](https:\u002F\u002Fdevelopers.google.com\u002Fmachine-learning\u002Fguides\u002Ftext-classification\u002F)\n1. [Deep Learning for NLP with PyTorch](https:\u002F\u002Fpytorch.org\u002Ftutorials\u002Fbeginner\u002Fdeep_learning_nlp_tutorial.html)\n\nTalks\n----\n1. [Deep Learning for Natural Language Processing (without Magic)](http:\u002F\u002Fwww.socher.org\u002Findex.php\u002FDeepLearningTutorial\u002FDeepLearningTutorial)\n1. [A Primer on Neural Network Models for Natural Language Processing](https:\u002F\u002Farxiv.org\u002Fabs\u002F1510.00726)\n1. [Deep Learning for Natural Language Processing: Theory and Practice (Tutorial)](https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002Fresearch\u002Fpublication\u002Fdeep-learning-for-natural-language-processing-theory-and-practice-tutorial\u002F)\n1. [TensorFlow Tutorials](https:\u002F\u002Fwww.tensorflow.org\u002Ftutorials\u002Fmandelbrot)\n1. [Practical Neural Networks for NLP](https:\u002F\u002Fgithub.com\u002Fclab\u002Fdynet_tutorial_examples) from EMNLP 2016 using DyNet framework\n1. [Recurrent Neural Networks with Word Embeddings](http:\u002F\u002Fdeeplearning.net\u002Ftutorial\u002Frnnslu.html)\n1. [LSTM Networks for Sentiment Analysis](http:\u002F\u002Fdeeplearning.net\u002Ftutorial\u002Flstm.html)\n1. [TensorFlow demo using the Large Movie Review Dataset](http:\u002F\u002Fai.stanford.edu\u002F~amaas\u002Fdata\u002Fsentiment\u002F)\n1. [LSTMVis: Visual Analysis for Recurrent Neural Networks](http:\u002F\u002Flstm.seas.harvard.edu\u002Fclient\u002Findex.html)\n1. Using deep learning in natural language processing by Rob Romijnders from PyData Amsterdam 2017\n\t- [video](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=HVdPWoZ_swY)\n\t- [slides](https:\u002F\u002Fgithub.com\u002FRobRomijnders\u002Ftalks\u002Fblob\u002Fmaster\u002Fpydata_DL_NLP.pdf)\n1. [Richard Socher's talk on sentiment analysis, question answering, and sentence-image embeddings](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=tdLmf8t4oqM)\n1. [Deep Learning, an interactive introduction for NLP-ers](http:\u002F\u002Fwww.slideshare.net\u002Froelofp\u002F220115dlmeetup)\n1. [Deep Natural Language Understanding](http:\u002F\u002Fvideolectures.net\u002Fdeeplearning2016_cho_language_understanding\u002F)\n1. [Deep Learning Summer School, Montreal 2016](http:\u002F\u002Fvideolectures.net\u002Fdeeplearning2016_montreal\u002F) Includes state-of-art language modeling.\n1. Tackling the Limits of Deep Learning for NLP by Richard Socher\n\t- [video](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=JYwNmSe4HqE)\n\t- [slides](https:\u002F\u002Fberkeley-deep-learning.github.io\u002Fcs294-131-s17\u002Fslides\u002Fsocher-talk.pdf)\n\nFrameworks\n----\n1. [Overview of DL frameworks for NLP](https:\u002F\u002Fmedium.com\u002F@datamonsters\u002F13-deep-learning-frameworks-for-natural-language-processing-in-python-2b84a6b6cd98)\n1. General Frameworks\n\t1. [Keras](https:\u002F\u002Fkeras.io\u002F) - _The Python Deep Learning library_ Emphasis on user friendliness, modularity, easy extensibility, and Pythonic.\n\t1. [TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002F) - A cross-platform, general purpose Machine Intelligence library with Python and C++ API.\n\t1. [PyTorch](http:\u002F\u002Fpytorch.org\u002F) - PyTorch is a deep learning framework that puts Python first. \"Tensors and Dynamic neural networks in Python with strong GPU acceleration.\"\n\n1. Specific Frameworks\n\t1. [SpaCy](https:\u002F\u002Fspacy.io\u002F) - A Python package designed for speed, getting things dones, and interoperates with other Deep Learning frameworks\n\t1. [Genism: Topic modeling for humans](https:\u002F\u002Fpypi.python.org\u002Fpypi\u002Fgensim) - A Python package that includes word2vec and doc2vec implementations.\n\t1. [fasttext](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002FfastText) Facebook's library for fast text representation and classification.\n\t1. Built on TensorFlow\n\t\t1. [SyntaxNet](https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Fmodels\u002Ftree\u002Fmaster\u002Fresearch\u002Fsyntaxnet) - A toolkit for natural language understanding (NLU).\n\t\t1. [textsum](https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Fmodels\u002Ftree\u002Fmaster\u002Fresearch\u002Ftextsum) - A Sequence-to-Sequence with Attention Model for Text Summarization.\n\t\t1. [Skip-Thought Vectors](https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Fmodels\u002Ftree\u002Fmaster\u002Fresearch\u002Fskip_thoughts) implementation in TensorFlow.\n\t\t1. [ActiveQA: Active Question Answering](https:\u002F\u002Fgithub.com\u002Fgoogle\u002Factive-qa) - Using reinforcement learning to train artificial agents for question answering\n\t\t1. [BERT](https:\u002F\u002Fgithub.com\u002Fgoogle-research\u002Fbert) - Bidirectional Encoder Representations from Transformers for pre-trained models\n\t1. Built on PyTorch\n\t\t1. [PyText](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fpytext) - A deep-learning based NLP modeling framework by Facebook\n\t\t1. [AllenNLP](https:\u002F\u002Fallennlp.org\u002F) - An open-source NLP research library\n\t\t1. [Flair](https:\u002F\u002Fgithub.com\u002Fzalandoresearch\u002Fflair) - A very simple framework for state-of-the-art NLP\n\t\t1. [fairseq](https:\u002F\u002Fgithub.com\u002Fpytorch\u002Ffairseq) - A Sequence-to-Sequence Toolkit\n\t\t1. [fastai](http:\u002F\u002Fdocs.fast.ai\u002Ftext.html) - Simplifies training fast and accurate neural nets using modern best practices\n\t\t1. [Transformer model](http:\u002F\u002Fnlp.seas.harvard.edu\u002F2018\u002F04\u002F03\u002Fattention.html) - Annotated notebook implementation\n\t1. [Deeplearning4j’s NLP framework](http:\u002F\u002Fdeeplearning4j.org\u002Fnlp) - Java implementation.\n\t1. [DyNet](https:\u002F\u002Fgithub.com\u002Fclab\u002Fdynet) - _The Dynamic Neural Network Toolkit_ \"work well with networks that have dynamic structures that change for every training instance\".\n\t1. [deepnl](https:\u002F\u002Fgithub.com\u002Fattardi\u002Fdeepnl) - A Python library for NLP based on Deep Learning neural network architecture.\n\nPapers\n----\n1. [Deep or shallow, NLP is breaking out](http:\u002F\u002Fdl.acm.org\u002Fcitation.cfm?id=2874915) - General overview of how Deep Learning is impacting NLP.\n1. [Natural Language Processing from Research at Google](http:\u002F\u002Fresearch.google.com\u002Fpubs\u002FNaturalLanguageProcessing.html) - Not all Deep Learning (but mostly).\n1. [Context Dependent Recurrent Neural Network Language Model](http:\u002F\u002Fwww.msr-waypoint.com\u002Fpubs\u002F176926\u002Frnn_ctxt.pdf)\n1. [Translation Modeling with Bidirectional Recurrent Neural Networks](https:\u002F\u002Fwww-i6.informatik.rwth-aachen.de\u002Fpublications\u002Fdownload\u002F936\u002FSundermeyerMartinAlkhouliTamerWuebkerJoernNeyHermann--TranslationModelingwithBidirectionalRecurrentNeuralNetworks--2014.pdf)\n1. [Contextual LSTM (CLSTM) models for Large scale NLP tasks](https:\u002F\u002Farxiv.org\u002Fabs\u002F1602.06291)\n1. [LSTM Neural Networks for Language Modeling](http:\u002F\u002Fciteseerx.ist.psu.edu\u002Fviewdoc\u002Fdownload?doi=10.1.1.248.4448&rep=rep1&type=pdf)\n1. [Exploring the Limits of Language Modeling](http:\u002F\u002Farxiv.org\u002Fpdf\u002F1602.02410.pdf)\n1. [Conversational Contextual Cues](https:\u002F\u002Farxiv.org\u002Fabs\u002F1606.00372) - Models context and participants in conversations.\n1. [Sequence to sequence learning with neural networks](http:\u002F\u002Fpapers.nips.cc\u002Fpaper\u002F5346-sequence-to-sequence-learning-with-neural-networks.pdf)\n1. [Efficient Estimation of Word Representations in Vector Space](http:\u002F\u002Farxiv.org\u002Fpdf\u002F1301.3781.pdf)\n1. [Learning Character-level Representations for Part-of-Speech Tagging](http:\u002F\u002Fjmlr.org\u002Fproceedings\u002Fpapers\u002Fv32\u002Fsantos14.pdf)\n1. [Representation Learning for Text-level Discourse Parsing](http:\u002F\u002Fwww.cc.gatech.edu\u002F~jeisenst\u002Fpapers\u002Fji-acl-2014.pdf)\n1. [Fast and Robust Neural Network Joint Models for Statistical Machine Translation](http:\u002F\u002Facl2014.org\u002Facl2014\u002FP14-1\u002Fpdf\u002FP14-1129.pdf)\n1. [Parsing With Compositional Vector Grammars](http:\u002F\u002Fwww.socher.org\u002Findex.php\u002FMain\u002FParsingWithCompositionalVectorGrammars)\n1. [Smart Reply: Automated Response Suggestion for Email](https:\u002F\u002Farxiv.org\u002Fabs\u002F1606.04870)\n1. [Neural Architectures for Named Entity Recognition](https:\u002F\u002Farxiv.org\u002Fabs\u002F1603.01360) - State-of-the-art performance in NER with bidirectional LSTM with a sequential conditional random layer and transition-based parsing with stack LSTMs.\n1. [Grammar as a Foreign Language](https:\u002F\u002Farxiv.org\u002Fabs\u002F1412.7449) - State-of-the-art syntactic constituency parsing using generic sequence-to-sequence approach.\n\nBlog Posts\n----\n\n1. [Natural Language Processing (NLP) progress](https:\u002F\u002Fnlpprogress.com\u002F) Tracking the most common NLP tasks, including the datasets and the current state-of-the-art \n1. [A Review of the Recent History of Natural Language Processing](http:\u002F\u002Fblog.aylien.com\u002Fa-review-of-the-recent-history-of-natural-language-processing\u002F)\n1. [Deep Learning, NLP, and Representations](http:\u002F\u002Fcolah.github.io\u002Fposts\u002F2014-07-NLP-RNNs-Representations\u002F)\n1. [The Unreasonable Effectiveness of Recurrent Neural Networks](http:\u002F\u002Fkarpathy.github.io\u002F2015\u002F05\u002F21\u002Frnn-effectiveness\u002F)\n1. [Neural Language Modeling From Scratch](http:\u002F\u002Fofir.io\u002FNeural-Language-Modeling-From-Scratch\u002F?a=1)\n1. [Machine Learning for Emoji Trends](http:\u002F\u002Finstagram-engineering.tumblr.com\u002Fpost\u002F117889701472\u002Femojineering-part-1-machine-learning-for-emoji)\n1. [Teaching Robots to Feel: Emoji & Deep Learning](http:\u002F\u002Fgetdango.com\u002Femoji-and-deep-learning.html)\n1. [Computational Linguistics and Deep Learning](http:\u002F\u002Fwww.mitpressjournals.org\u002Fdoi\u002Fpdf\u002F10.1162\u002FCOLI_a_00239) - Opinion piece on how Deep Learning fits into the broader picture of text processing.\n1. [Deep Learning NLP Best Practices](http:\u002F\u002Fruder.io\u002Fdeep-learning-nlp-best-practices\u002Findex.html)\n1. [7 types of Artificial Neural Networks for Natural Language Processing](https:\u002F\u002Fmedium.com\u002F@datamonsters\u002Fartificial-neural-networks-for-natural-language-processing-part-1-64ca9ebfa3b2)\n1. [How to solve 90% of NLP problems: a step-by-step guide](https:\u002F\u002Fblog.insightdatascience.com\u002Fhow-to-solve-90-of-nlp-problems-a-step-by-step-guide-fda605278e4e)\n2. [7 Applications of Deep Learning for Natural Language Processing](https:\u002F\u002Fmachinelearningmastery.com\u002Fapplications-of-deep-learning-for-natural-language-processing\u002F)\n\nDatasets\n----\n1. [Dataset from \"One Billion Word Language Modeling Benchmark\"](http:\u002F\u002Fwww.statmt.org\u002Flm-benchmark\u002F1-billion-word-language-modeling-benchmark-r13output.tar.gz) - Almost 1B words, already pre-processed text.\n1. [Stanford Sentiment Treebank](https:\u002F\u002Fnlp.stanford.edu\u002Fsentiment\u002Ftreebank.html) - Fine grained sentiment labels for 215,154 phrases in the parse trees of 11,855 sentences.\n1. [Chatbot data from Kaggle](https:\u002F\u002Fwww.kaggle.com\u002Fsamdeeplearning\u002Fdeepnlp)\n1. [A list of text datasets that are free\u002Fpublic domain in alphabetical order](https:\u002F\u002Fgithub.com\u002Fniderhoff\u002Fnlp-datasets)\n1. [Another list of text datasets that are free\u002Fpublic domain in reverse chronological order](https:\u002F\u002Fgithub.com\u002Fkarthikncode\u002Fnlp-datasets)\n1. Question Answering datasets\n\t1. [Quora's Question Pairs Dataset](https:\u002F\u002Fdata.quora.com\u002FFirst-Quora-Dataset-Release-Question-Pairs) - Identify question pairs that have the same intent.\n\t1. [CMU's Wikipedia Factoid Question Answers](https:\u002F\u002Fwww.cs.cmu.edu\u002F~ark\u002FQA-data\u002F)\n\t1. [DeepMind's Algebra Question Answering](https:\u002F\u002Fgithub.com\u002Fdeepmind\u002FAQuA)\n\t1. [DeepMind's from CNN & DailyMail Question Answering](https:\u002F\u002Fgithub.com\u002Fdeepmind\u002Frc-data)\n\t1. [Microsoft's WikiQA Open Domain Question Answering](https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002Fresearch\u002Fpublication\u002Fwikiqa-a-challenge-dataset-for-open-domain-question-answering\u002F)\n\t1. [Stanford Question Answering Dataset (SQuAD)](https:\u002F\u002Frajpurkar.github.io\u002FSQuAD-explorer\u002F) - covering reading comprehension\n\nWord Embeddings and friends\n----\n1. [The amazing power of word vectors](https:\u002F\u002Fblog.acolyer.org\u002F2016\u002F04\u002F21\u002Fthe-amazing-power-of-word-vectors\u002F) from The Morning Paper blog\n1. [Distributed Representations of Words and Phrases and their Compositionality](https:\u002F\u002Fpapers.nips.cc\u002Fpaper\u002F5021-distributed-representations-of-words-and-phrases-and-their-compositionality.pdf) - The original word2vec paper.\n1. [word2vec Parameter Learning Explained](https:\u002F\u002Farxiv.org\u002Fabs\u002F1411.2738) An elucidating explanation of word2vec training\n1. [Word embeddings in 2017: Trends and future directions](http:\u002F\u002Fruder.io\u002Fword-embeddings-2017\u002F)\n1. [Learning Word Vectors for 157 Languages](https:\u002F\u002Farxiv.org\u002Fabs\u002F1802.06893)\n1. [GloVe: Global Vectors for Word Representation](http:\u002F\u002Fwww-nlp.stanford.edu\u002Fpubs\u002Fglove.pdf) - A \"count-based\"\u002Fco-occurrence model to learn word embeddings.\n1.  Doc2Vec\n\t- [A gentle introduction to Doc2Vec](https:\u002F\u002Fmedium.com\u002Fscaleabout\u002Fa-gentle-introduction-to-doc2vec-db3e8c0cce5e)\n\t- [Distributed Representations of Sentences and Documents](https:\u002F\u002Fcs.stanford.edu\u002F~quocle\u002Fparagraph_vector.pdf)\n1. [Dynamic word embeddings for evolving semantic discovery](https:\u002F\u002Fblog.acolyer.org\u002F2018\u002F02\u002F22\u002Fdynamic-word-embeddings-for-evolving-semantic-discovery\u002F) from The Morning Paper blog\n1. Ali Ghodsi's lecture on word2vec: \n\t- [part 1](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=TsEGsdVJjuA)\n\t- [part 2](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=nuirUEmbaJU)\n1. [word2vec analogy demo](http:\u002F\u002Fdeeplearner.fz-qqq.net\u002F)\n1. [TensorFlow Embedding Projector of word vectors](http:\u002F\u002Fprojector.tensorflow.org\u002F)\n1. Skip-Thought Vectors - \"unsupervised learning of a generic, distributed sentence encoder\"\n    - [Paper](http:\u002F\u002Farxiv.org\u002Fabs\u002F1506.06726)\n    - [Code](https:\u002F\u002Fgithub.com\u002Fryankiros\u002Fskip-thoughts)\n\n-----\nContributing\n----\nHave anything in mind that you think is awesome and would fit in this list? Feel free to send me a pull request!\n\n-----\nLicense\n----\n\n[![CC0](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fbrianspiering_awesome-dl4nlp_readme_b7657951a0bb.png)](http:\u002F\u002Fcreativecommons.org\u002Fpublicdomain\u002Fzero\u002F1.0\u002F)\n\nTo the extent possible under law, [Dr. Brian J. Spiering](http:\u002F\u002Fwww.linkedin.com\u002Fin\u002Fbrianspiering\u002F) has waived all copyright and related or neighboring rights to this work.\n","自然语言处理（NLP）中的优秀深度学习资源 [![Awesome](https:\u002F\u002Fcdn.rawgit.com\u002Fsindresorhus\u002Fawesome\u002Fd7305f38d29fed78fa85652e3a63e154dd8e8829\u002Fmedia\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Fsindresorhus\u002Fawesome)\n====\n\n目录\n----\n\n- __[课程](#courses)__\n- __[书籍](#books)__\n- __[教程](#tutorials)__\n- __[演讲\u002F讲座](#talks)__\n- __[框架\u002F模型](#frameworks)__\n- __[论文](#papers)__\n- __[博客文章](#blog-posts)__\n- __[数据集](#datasets)__\n- __[词嵌入\u002F词向量](#word-embeddings)__\n- __[贡献](#contributing)__\n\n课程\n----\n1. 斯坦福大学深度学习与自然语言处理课程 \u002F CS224N（2019年冬季）\n\t- [课程主页](http:\u002F\u002Fweb.stanford.edu\u002Fclass\u002Fcs224n\u002Findex.html) 本课程提供了该领域的全面概述，包含视频、讲义以及学生项目示例。\n\t- [课程视频](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=8rXD5-xhemo&list=PLoROMvodv4rOhcuXMZkNm7j3fVwBBY42z) 视频播放列表。\n\t- [往期课程笔记](https:\u002F\u002Fgithub.com\u002Fstanfordnlp\u002Fcs224n-winter17-notes) 或许是关于NLP深度学习的最佳“教材”。\n\t- [课程代码](https:\u002F\u002Fgithub.com\u002FDSKSD\u002FDeepNLP-models-Pytorch) 包含CS224N中各种深度NLP模型的PyTorch实现。\n1. 卡内基梅隆大学的NLP神经网络课程\n\t- [课程主页](http:\u002F\u002Fphontron.com\u002Fclass\u002Fnn4nlp2017\u002F)\n\t- [课程视频](https:\u002F\u002Fwww.youtube.com\u002Fuser\u002Fneubig\u002Fvideos)\n\t- [课程代码](https:\u002F\u002Fgithub.com\u002Fneubig\u002Fnn4nlp2017-code\u002F)\n1. 牛津大学与DeepMind联合开设的自然语言处理深度学习课程\n\t- [课程主页](https:\u002F\u002Fwww.cs.ox.ac.uk\u002Fteaching\u002Fcourses\u002F2016-2017\u002Fdl\u002F)\n\t- [课程幻灯片](https:\u002F\u002Fgithub.com\u002Foxford-cs-deepnlp-2017\u002Flectures)\n\t- [课程视频](https:\u002F\u002Fwww.youtube.com\u002Fplaylist?list=PL613dYIGMXoZBtZhbyiBqb0QtgK6oJbpm)\n\n书籍\n-----\n1. [用文本进行深度学习：几乎从零开始的自然语言处理——Python与spaCy](https:\u002F\u002Fwww.amazon.com\u002FDeep-Learning-Text-Approach-Processing\u002Fdp\u002F1491984414)，作者：帕特里克·哈里森和马修·霍尼巴尔\n1. [自然语言处理中的神经网络方法](https:\u002F\u002Fwww.amazon.com\u002Fgp\u002Fproduct\u002F1627052984)，作者：约阿夫·戈德堡和格雷厄姆·赫斯特\n1. [自然语言处理中的深度学习](http:\u002F\u002Fwww.springer.com\u002Fus\u002Fbook\u002F9789811052088)，作者：邓莉和刘洋\n1. [人工智能背后的数学：AI基础指南【完整书籍】](https:\u002F\u002Fwww.freecodecamp.org\u002Fnews\u002Fthe-math-behind-artificial-intelligence-book\u002F)——一本免费书籍，以通俗易懂的英语解释并结合Python代码示例讲解AI背后的数学原理\n1. [实战自然语言处理](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fnatural-language-processing-in-action)，作者：霍布森·莱恩、科尔·霍华德和汉内斯·哈普克\n1. The LazyProgrammer所著的《Python中的自然语言处理深度学习》（仅限Kindle版）\n\t1. [Python与Theano中的Word2Vec及词嵌入](https:\u002F\u002Fwww.amazon.com\u002FDeep-Learning-Language-Processing-Embeddings-ebook\u002Fdp\u002FB01KQ0ZN0A)\n\t1. [Python与Theano中从Word2Vec到GLoVe](https:\u002F\u002Fwww.amazon.com\u002FDeep-Learning-Language-Processing-Word2Vec-ebook\u002Fdp\u002FB01KRBOO4Y\u002F)\n\t1. [递归神经网络：Theano中的递归神经（张量）网络](https:\u002F\u002Fwww.amazon.com\u002FDeep-Learning-Language-Processing-Recursive-ebook\u002Fdp\u002FB01KS5AEXO)\n1. [Python应用自然语言处理](https:\u002F\u002Fwww.amazon.ca\u002FApplied-Natural-Language-Processing-Python\u002Fdp\u002F1484237323)，作者：塔韦·贝索洛夫二世\n1. [深度学习烹饪书](https:\u002F\u002Fwww.amazon.ca\u002FDeep-Learning-Cookbook-Practical-Recipes\u002Fdp\u002F149199584X)，作者：杜威·奥辛加\n1. [自然语言处理深度学习：使用Python构建神经网络](https:\u002F\u002Fwww.amazon.ca\u002FDeep-Learning-Natural-Language-Processing\u002Fdp\u002F148423684X)，作者：帕拉什·戈亚尔、苏米特·潘迪和卡兰·贾因\n1. [文本的机器学习](https:\u002F\u002Fwww.amazon.ca\u002FMachine-Learning-Text-Charu-Aggarwal\u002Fdp\u002F3319735306)，作者：查鲁·C·阿加瓦尔\n1. [TensorFlow自然语言处理](https:\u002F\u002Fwww.amazon.ca\u002FNatural-Language-Processing-TensorFlow-language-ebook\u002Fdp\u002FB077Q3VZFR)，作者：图山·加内格达拉\n1. [fastText快速入门指南：开始使用Facebook的文本表示与分类库](https:\u002F\u002Fwww.amazon.ca\u002FfastText-Quick-Start-Guide-representation\u002Fdp\u002F1789130999)\n1. [动手实践Python自然语言处理](https:\u002F\u002Fwww.amazon.ca\u002FHands-Natural-Language-Processing-Python\u002Fdp\u002F178913949X)\n1. [实战自然语言处理，第二版](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fnatural-language-processing-in-action-second-edition)，作者：霍布森·莱恩和玛丽亚·迪谢尔\n1. [实战自然语言处理入门](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fgetting-started-with-natural-language-processing)，作者：叶卡捷琳娜·科奇马尔\n2. [实战自然语言处理深度学习](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fdeep-learning-for-natural-language-processing)，作者：斯蒂芬·拉伊马克尔斯\n\n教程\n-----\n\n1. [谷歌文本分类指南](https:\u002F\u002Fdevelopers.google.com\u002Fmachine-learning\u002Fguides\u002Ftext-classification\u002F)\n1. [使用PyTorch进行NLP深度学习](https:\u002F\u002Fpytorch.org\u002Ftutorials\u002Fbeginner\u002Fdeep_learning_nlp_tutorial.html)\n\n讲座\n----\n1. [自然语言处理中的深度学习（无需魔法）](http:\u002F\u002Fwww.socher.org\u002Findex.php\u002FDeepLearningTutorial\u002FDeepLearningTutorial)\n1. [自然语言处理神经网络模型入门](https:\u002F\u002Farxiv.org\u002Fabs\u002F1510.00726)\n1. [自然语言处理中的深度学习：理论与实践（教程）](https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002Fresearch\u002Fpublication\u002Fdeep-learning-for-natural-language-processing-theory-and-practice-tutorial\u002F)\n1. [TensorFlow 教程](https:\u002F\u002Fwww.tensorflow.org\u002Ftutorials\u002Fmandelbrot)\n1. 使用 DyNet 框架的 EMNLP 2016 实践性 NLP 神经网络教程\n1. [带有词嵌入的循环神经网络](http:\u002F\u002Fdeeplearning.net\u002Ftutorial\u002Frnnslu.html)\n1. [用于情感分析的 LSTM 网络](http:\u002F\u002Fdeeplearning.net\u002Ftutorial\u002Flstm.html)\n1. 使用大型电影评论数据集的 TensorFlow 演示\n1. [LSTMVis：针对循环神经网络的可视化分析](http:\u002F\u002Flstm.seas.harvard.edu\u002Fclient\u002Findex.html)\n1. 来自 PyData 阿姆斯特丹 2017 的 Rob Romijnders 关于在自然语言处理中使用深度学习的演讲\n\t- [视频](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=HVdPWoZ_swY)\n\t- [幻灯片](https:\u002F\u002Fgithub.com\u002FRobRomijnders\u002Ftalks\u002Fblob\u002Fmaster\u002Fpydata_DL_NLP.pdf)\n1. [Richard Socher 关于情感分析、问答和句子-图像嵌入的演讲](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=tdLmf8t4oqM)\n1. [深度学习：面向 NLP 从业者的互动式入门](http:\u002F\u002Fwww.slideshare.net\u002Froelofp\u002F220115dlmeetup)\n1. [深度自然语言理解](http:\u002F\u002Fvideolectures.net\u002Fdeeplearning2016_cho_language_understanding\u002F)\n1. [2016 年蒙特利尔深度学习暑期学校](http:\u002F\u002Fvideolectures.net\u002Fdeeplearning2016_montreal\u002F) 包括最先进的语言建模。\n1. Richard Socher 的“应对深度学习在 NLP 中的局限性”演讲\n\t- [视频](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=JYwNmSe4HqE)\n\t- [幻灯片](https:\u002F\u002Fberkeley-deep-learning.github.io\u002Fcs294-131-s17\u002Fslides\u002Fsocher-talk.pdf)\n\n框架\n----\n1. [NLP 领域的深度学习框架概述](https:\u002F\u002Fmedium.com\u002F@datamonsters\u002F13-deep-learning-frameworks-for-natural-language-processing-in-python-2b84a6b6cd98)\n1. 通用框架\n\t1. [Keras](https:\u002F\u002Fkeras.io\u002F) - _Python 深度学习库_ 强调用户友好性、模块化、易于扩展以及 Python 式的设计。\n\t1. [TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002F) - 跨平台、通用的机器智能库，提供 Python 和 C++ API。\n\t1. [PyTorch](http:\u002F\u002Fpytorch.org\u002F) - PyTorch 是一个以 Python 为先的深度学习框架。“在 Python 中实现张量和动态神经网络，并具有强大的 GPU 加速功能。”\n\n1. 特定框架\n\t1. [SpaCy](https:\u002F\u002Fspacy.io\u002F) - 专为速度、高效完成任务而设计的 Python 包，可与其他深度学习框架互操作。\n\t1. [Genism：面向人类的主题建模](https:\u002F\u002Fpypi.python.org\u002Fpypi\u002Fgensim) - 一个包含 word2vec 和 doc2vec 实现的 Python 包。\n\t1. [fasttext](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002FfastText) Facebook 开发的快速文本表示与分类库。\n\t1. 基于 TensorFlow 构建\n\t\t1. [SyntaxNet](https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Fmodels\u002Ftree\u002Fmaster\u002Fresearch\u002Fsyntaxnet) - 用于自然语言理解（NLU）的工具包。\n\t\t1. [textsum](https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Fmodels\u002Ftree\u002Fmaster\u002Fresearch\u002Ftextsum) - 一种带有注意力机制的序列到序列文本摘要模型。\n\t\t1. [Skip-Thought Vectors](https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Fmodels\u002Ftree\u002Fmaster\u002Fresearch\u002Fskip_thoughts) 的 TensorFlow 实现。\n\t\t1. [ActiveQA：主动问答](https:\u002F\u002Fgithub.com\u002Fgoogle\u002Factive-qa) - 利用强化学习训练人工代理进行问答。\n\t\t1. [BERT](https:\u002F\u002Fgithub.com\u002Fgoogle-research\u002Fbert) - 基于 Transformer 的双向编码器表示预训练模型。\n\t1. 基于 PyTorch 构建\n\t\t1. [PyText](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fpytext) - Facebook 开发的基于深度学习的 NLP 建模框架。\n\t\t1. [AllenNLP](https:\u002F\u002Fallennlp.org\u002F) - 一个开源的 NLP 研究库。\n\t\t1. [Flair](https:\u002F\u002Fgithub.com\u002Fzalandoresearch\u002Fflair) - 一个非常简单的框架，用于实现最先进的 NLP 技术。\n\t\t1. [fairseq](https:\u002F\u002Fgithub.com\u002Fpytorch\u002Ffairseq) - 一个序列到序列工具包。\n\t\t1. [fastai](http:\u002F\u002Fdocs.fast.ai\u002Ftext.html) - 通过现代最佳实践简化快速且准确的神经网络训练。\n\t\t1. [Transformer 模型](http:\u002F\u002Fnlp.seas.harvard.edu\u002F2018\u002F04\u002F03\u002Fattention.html) - 注释版笔记本实现。\n\t1. [Deeplearning4j 的 NLP 框架](http:\u002F\u002Fdeeplearning4j.org\u002Fnlp) - Java 实现。\n\t1. [DyNet](https:\u002F\u002Fgithub.com\u002Fclab\u002Fdynet) - _动态神经网络工具包_ “非常适合处理每轮训练都会变化的动态结构网络”。\n\t1. [deepnl](https:\u002F\u002Fgithub.com\u002Fattardi\u002Fdeepnl) - 一个基于深度学习神经网络架构的 Python NLP 库。\n\n论文\n----\n1. [深度还是浅层，NLP正迎来爆发](http:\u002F\u002Fdl.acm.org\u002Fcitation.cfm?id=2874915) - 深度学习如何影响自然语言处理的总体概述。\n1. [谷歌研究中的自然语言处理](http:\u002F\u002Fresearch.google.com\u002Fpubs\u002FNaturalLanguageProcessing.html) - 并非全部基于深度学习，但大部分是。\n1. [上下文相关的循环神经网络语言模型](http:\u002F\u002Fwww.msr-waypoint.com\u002Fpubs\u002F176926\u002Frnn_ctxt.pdf)\n1. [基于双向循环神经网络的翻译建模](https:\u002F\u002Fwww-i6.informatik.rwth-aachen.de\u002Fpublications\u002Fdownload\u002F936\u002FSundermeyerMartinAlkhouliTamerWuebkerJoernNeyHermann--TranslationModelingwithBidirectionalRecurrentNeuralNetworks--2014.pdf)\n1. [用于大规模NLP任务的上下文LSTM（CLSTM）模型](https:\u002F\u002Farxiv.org\u002Fabs\u002F1602.06291)\n1. [用于语言建模的LSTM神经网络](http:\u002F\u002Fciteseerx.ist.psu.edu\u002Fviewdoc\u002Fdownload?doi=10.1.1.248.4448&rep=rep1&type=pdf)\n1. [探索语言建模的极限](http:\u002F\u002Farxiv.org\u002Fpdf\u002F1602.02410.pdf)\n1. [对话中的上下文线索](https:\u002F\u002Farxiv.org\u002Fabs\u002F1606.00372) - 对话中的上下文和参与者进行建模。\n1. [基于神经网络的序列到序列学习](http:\u002F\u002Fpapers.nips.cc\u002Fpaper\u002F5346-sequence-to-sequence-learning-with-neural-networks.pdf)\n1. [向量空间中词表示的有效估计](http:\u002F\u002Farxiv.org\u002Fpdf\u002F1301.3781.pdf)\n1. [学习词级表示以进行词性标注](http:\u002F\u002Fjmlr.org\u002Fproceedings\u002Fpapers\u002Fv32\u002Fsantos14.pdf)\n1. [文本级话语解析的表示学习](http:\u002F\u002Fwww.cc.gatech.edu\u002F~jeisenst\u002Fpapers\u002Fji-acl-2014.pdf)\n1. [用于统计机器翻译的高效稳健的神经网络联合模型](http:\u002F\u002Facl2014.org\u002Facl2014\u002FP14-1\u002Fpdf\u002FP14-1129.pdf)\n1. [使用组合向量文法进行句法分析](http:\u002F\u002Fwww.socher.org\u002Findex.php\u002FMain\u002FParsingWithCompositionalVectorGrammars)\n1. [智能回复：电子邮件自动回复建议](https:\u002F\u002Farxiv.org\u002Fabs\u002F1606.04870)\n1. [命名实体识别的神经架构](https:\u002F\u002Farxiv.org\u002Fabs\u002F1603.01360) - 使用双向LSTM结合顺序条件随机场层以及基于栈的LSTM转换解析，在NER任务上达到最先进水平。\n1. [语法作为一门外语](https:\u002F\u002Farxiv.org\u002Fabs\u002F1412.7449) - 采用通用的序列到序列方法实现最先进的句法成分解析。\n\n博客文章\n----\n\n1. [自然语言处理（NLP）进展](https:\u002F\u002Fnlpprogress.com\u002F) 跟踪最常见的NLP任务，包括数据集和当前最先进水平。\n1. [自然语言处理近期历史回顾](http:\u002F\u002Fblog.aylien.com\u002Fa-review-of-the-recent-history-of-natural-language-processing\u002F)\n1. [深度学习、NLP与表示](http:\u002F\u002Fcolah.github.io\u002Fposts\u002F2014-07-NLP-RNNs-Representations\u002F)\n1. [循环神经网络的不可思议有效性](http:\u002F\u002Fkarpathy.github.io\u002F2015\u002F05\u002F21\u002Frnn-effectiveness\u002F)\n1. [从零开始的神经语言建模](http:\u002F\u002Fofir.io\u002FNeural-Language-Modeling-From-Scratch\u002F?a=1)\n1. [机器学习在表情符号趋势中的应用](http:\u002F\u002Finstagram-engineering.tumblr.com\u002Fpost\u002F117889701472\u002Femojineering-part-1-machine-learning-for-emoji)\n1. [教机器人感受：表情符号与深度学习](http:\u002F\u002Fgetdango.com\u002Femoji-and-deep-learning.html)\n1. [计算语言学与深度学习](http:\u002F\u002Fwww.mitpressjournals.org\u002Fdoi\u002Fpdf\u002F10.1162\u002FCOLI_a_00239) - 关于深度学习如何融入更广泛的文本处理领域的评论文章。\n1. [深度学习NLP最佳实践](http:\u002F\u002Fruder.io\u002Fdeep-learning-nlp-best-practices\u002Findex.html)\n1. [自然语言处理的7种人工神经网络](https:\u002F\u002Fmedium.com\u002F@datamonsters\u002Fartificial-neural-networks-for-natural-language-processing-part-1-64ca9ebfa3b2)\n1. [如何解决90%的NLP问题：分步指南](https:\u002F\u002Fblog.insightdatascience.com\u002Fhow-to-solve-90-of-nlp-problems-a-step-by-step-guide-fda605278e4e)\n2. [深度学习在自然语言处理中的7大应用](https:\u002F\u002Fmachinelearningmastery.com\u002Fapplications-of-deep-learning-for-natural-language-processing\u002F)\n\n数据集\n----\n1. [“十亿词语言建模基准”数据集](http:\u002F\u002Fwww.statmt.org\u002Flm-benchmark\u002F1-billion-word-language-modeling-benchmark-r13output.tar.gz) - 接近10亿词，已预处理过的文本。\n1. [斯坦福情感树库](https:\u002F\u002Fnlp.stanford.edu\u002Fsentiment\u002Ftreebank.html) - 为11,855个句子的句法树中的215,154个短语提供细粒度的情感标签。\n1. [Kaggle上的聊天机器人数据](https:\u002F\u002Fwww.kaggle.com\u002Fsamdeeplearning\u002Fdeepnlp)\n1. [按字母顺序排列的免费\u002F公共领域文本数据集列表](https:\u002F\u002Fgithub.com\u002Fniderhoff\u002Fnlp-datasets)\n1. [按时间倒序排列的另一份免费\u002F公共领域文本数据集列表](https:\u002F\u002Fgithub.com\u002Fkarthikncode\u002Fnlp-datasets)\n1. 问答数据集\n\t1. [Quora的问题对数据集](https:\u002F\u002Fdata.quora.com\u002FFirst-Quora-Dataset-Release-Question-Pairs) - 用于识别具有相同意图的问题对。\n\t1. [卡内基梅隆大学的维基百科事实型问答数据](https:\u002F\u002Fwww.cs.cmu.edu\u002F~ark\u002FQA-data\u002F)\n\t1. [DeepMind的代数问题解答](https:\u002F\u002Fgithub.com\u002Fdeepmind\u002FAQuA)\n\t1. [DeepMind来自CNN和DailyMail的问答数据](https:\u002F\u002Fgithub.com\u002Fdeepmind\u002Frc-data)\n\t1. [微软的WikiQA开放域问答数据](https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002Fresearch\u002Fpublication\u002Fwikiqa-a-challenge-dataset-for-open-domain-question-answering\u002F)\n\t1. [斯坦福问答数据集（SQuAD）](https:\u002F\u002Frajpurkar.github.io\u002FSQuAD-explorer\u002F) - 涵盖阅读理解。\n\n词嵌入及相关内容\n----\n1. 来自《The Morning Paper》博客的[词向量的惊人力量](https:\u002F\u002Fblog.acolyer.org\u002F2016\u002F04\u002F21\u002Fthe-amazing-power-of-word-vectors\u002F)\n1. [单词与短语的分布式表示及其组合性](https:\u002F\u002Fpapers.nips.cc\u002Fpaper\u002F5021-distributed-representations-of-words-and-phrases-and-their-compositionality.pdf) - 原始的word2vec论文。\n1. [word2vec参数学习详解](https:\u002F\u002Farxiv.org\u002Fabs\u002F1411.2738) 对word2vec训练过程的深入解释\n1. [2017年的词嵌入：趋势与未来方向](http:\u002F\u002Fruder.io\u002Fword-embeddings-2017\u002F)\n1. [为157种语言学习词向量](https:\u002F\u002Farxiv.org\u002Fabs\u002F1802.06893)\n1. [GloVe：用于词表示的全局向量](http:\u002F\u002Fwww-nlp.stanford.edu\u002Fpubs\u002Fglove.pdf) - 一种基于计数\u002F共现的模型，用于学习词嵌入。\n1. Doc2Vec\n\t- [Doc2Vec简明介绍](https:\u002F\u002Fmedium.com\u002Fscaleabout\u002Fa-gentle-introduction-to-doc2vec-db3e8c0cce5e)\n\t- [句子与文档的分布式表示](https:\u002F\u002Fcs.stanford.edu\u002F~quocle\u002Fparagraph_vector.pdf)\n1. 来自《The Morning Paper》博客的[用于动态语义发现的动态词嵌入](https:\u002F\u002Fblog.acolyer.org\u002F2018\u002F02\u002F22\u002Fdynamic-word-embeddings-for-evolving-semantic-discovery\u002F)\n1. Ali Ghodsi关于word2vec的讲座：\n\t- [第一部分](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=TsEGsdVJjuA)\n\t- [第二部分](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=nuirUEmbaJU)\n1. [word2vec类比演示](http:\u002F\u002Fdeeplearner.fz-qqq.net\u002F)\n1. [TensorFlow词向量投影仪](http:\u002F\u002Fprojector.tensorflow.org\u002F)\n1. Skip-Thought向量 - “无监督学习通用的分布式句子编码器”\n    - [论文](http:\u002F\u002Farxiv.org\u002Fabs\u002F1506.06726)\n    - [代码](https:\u002F\u002Fgithub.com\u002Fryankiros\u002Fskip-thoughts)\n\n-----\n贡献\n----\n你有没有什么觉得很棒、适合加入本列表的内容？欢迎随时向我提交拉取请求！\n\n-----\n许可\n----\n\n[![CC0](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fbrianspiering_awesome-dl4nlp_readme_b7657951a0bb.png)](http:\u002F\u002Fcreativecommons.org\u002Fpublicdomain\u002Fzero\u002F1.0\u002F)\n\n在法律允许的最大范围内，[Brian J. Spiering博士](http:\u002F\u002Fwww.linkedin.com\u002Fin\u002Fbrianspiering\u002F)已放弃本作品的所有版权及相关或邻接权利。","# awesome-dl4nlp 快速上手指南\n\n`awesome-dl4nlp` 并非一个可直接安装的单一软件库，而是一个精选的**深度学习与自然语言处理（NLP）资源清单**。它汇集了课程、书籍、教程、框架、论文和数据集。\n\n本指南将指导你如何利用该清单中的核心资源，快速搭建开发环境并运行第一个深度学习 NLP 项目。\n\n## 环境准备\n\n在开始之前，请确保你的系统满足以下要求。由于清单中大多数代码示例基于 **Python** 和主流深度学习框架（如 PyTorch 或 TensorFlow），建议按以下标准配置环境。\n\n### 系统要求\n- **操作系统**: Linux (推荐 Ubuntu 20.04+), macOS, 或 Windows (建议使用 WSL2)\n- **硬件**: 推荐使用带有 CUDA 支持的 NVIDIA GPU（显存 8GB+ 为佳），用于加速模型训练。若无 GPU，CPU 亦可运行小型示例。\n- **Python 版本**: 3.8 - 3.10\n\n### 前置依赖\n你需要安装 Python 包管理工具及基础科学计算库。\n\n```bash\n# 检查 Python 版本\npython3 --version\n\n# 安装 pip (如果尚未安装)\nsudo apt-get install python3-pip  # Linux\nbrew install python               # macOS\n\n# 推荐创建虚拟环境 (使用 venv 或 conda)\npython3 -m venv dl4nlp_env\nsource dl4nlp_env\u002Fbin\u002Factivate    # Windows 用户使用: dl4nlp_env\\Scripts\\activate\n```\n\n## 安装步骤\n\n由于 `awesome-dl4nlp` 是资源列表，我们选择清单中推荐的 **PyTorch** 框架配合 **Hugging Face Transformers**（现代 NLP 事实标准，源自清单中的 BERT\u002FTransformer 部分）作为入门起点。\n\n### 1. 配置国内镜像源（加速下载）\n为避免网络延迟，建议将 pip 源切换至清华大学或阿里云镜像。\n\n```bash\npip config set global.index-url https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n```\n\n### 2. 安装核心深度学习框架\n根据清单推荐，安装 PyTorch 及相关 NLP 库。\n\n```bash\n# 安装 PyTorch (CPU 版本，若需 GPU 请访问 pytorch.org 获取对应 CUDA 命令)\npip install torch torchvision torchaudio\n\n# 安装 NLP 专用库 (对应清单中的 Frameworks 部分)\npip install transformers datasets spacy scikit-learn\n\n# 下载 SpaCy 中文模型 (可选，用于预处理)\npython -m spacy download zh_core_web_sm\n```\n\n### 3. 获取示例代码\n你可以克隆清单中提到的斯坦福 CS224N 课程代码仓库作为学习起点：\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002FDSKSD\u002FDeepNLP-models-Pytorch.git\ncd DeepNLP-models-Pytorch\npip install -r requirements.txt\n```\n\n## 基本使用\n\n以下示例展示如何使用清单中推荐的 **Transformers** 库（基于 BERT 模型，见清单 Papers\u002FFrameworks 部分），在 5 行代码内完成中文文本情感分析。\n\n### 最简单的使用示例\n\n创建一个名为 `quick_start.py` 的文件：\n\n```python\nfrom transformers import pipeline\n\n# 1. 加载预训练的中文情感分析模型\n# 该模型对应清单中提到的 \"BERT\" 及 \"Pre-trained models\" 概念\nclassifier = pipeline(\"sentiment-analysis\", model=\"uer\u002Froberta-base-finetuned-jd-binary-chinese\")\n\n# 2. 准备测试文本\ntext = \"这款手机的屏幕非常清晰，但是电池续航能力太差了。\"\n\n# 3. 执行预测\nresult = classifier(text)\n\n# 4. 输出结果\nprint(f\"文本: {text}\")\nprint(f\"分析结果: {result}\")\n```\n\n### 运行代码\n\n```bash\npython quick_start.py\n```\n\n**预期输出：**\n模型将自动下载参数（首次运行需联网），并输出类似以下结果（标签可能为 NEGATIVE\u002FPOSITIVE 或 0\u002F1）：\n```text\n文本：这款手机的屏幕非常清晰，但是电池续航能力太差了。\n分析结果: [{'label': 'NEGATIVE', 'score': 0.998}]\n```\n\n### 下一步学习路径\n参考 `awesome-dl4nlp` 原始清单中的以下板块深入进阶：\n- **Courses**: 观看 Stanford CS224N 视频课程理解理论基础。\n- **Papers**: 阅读 \"Attention is All You Need\" (Transformer 原文) 或 BERT 相关论文。\n- **Frameworks**: 尝试使用 `AllenNLP` 或 `Flair` 构建更复杂的命名实体识别（NER）系统。","某初创公司的算法工程师小李正负责搭建一个智能客服系统，急需掌握最新的深度学习 NLP 技术以优化意图识别模块。\n\n### 没有 awesome-dl4nlp 时\n- **资源检索低效**：在谷歌和 GitHub 上盲目搜索\"Deep Learning NLP\"，结果充斥着过时教程或营销软文，难以甄别高质量内容。\n- **学习路径混乱**：面对零散的课程、书籍和论文，无法构建系统的知识体系，不知道是该先读斯坦福 CS224N 还是直接看代码实现。\n- **复现成本高昂**：寻找特定模型（如 Word2Vec 或递归神经网络）的开源代码时，常遇到文档缺失或依赖冲突的项目，浪费数天调试环境。\n- **前沿信息滞后**：缺乏权威渠道追踪最新的研究论文和技术博客，导致方案选型停留在两三年前的技术水平。\n\n### 使用 awesome-dl4nlp 后\n- **一站式获取精品**：直接访问分类清晰的清单，瞬间锁定斯坦福 CS224N 视频、CMU 课程代码及经典书籍，确保所有资源均经社区验证。\n- **结构化进阶学习**：依托其目录结构，按“课程 - 书籍 - 框架 - 论文”的顺序制定学习计划，从理论基础平滑过渡到实战演练。\n- **快速落地原型**：通过\"Frameworks \u002F Models\"板块直接找到基于 PyTorch 的高质量模型实现，将原本需要一周的环境搭建与代码复现缩短至几小时。\n- **紧跟技术前沿**：利用\"Papers\"和\"Blog Posts\"栏目实时跟踪最新研究成果，迅速将先进的注意力机制应用到客服系统中，提升识别准确率。\n\nawesome-dl4nlp 将原本碎片化且充满噪音的学习过程，转变为一条高效、系统且紧跟前沿的技术进阶快车道。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fbrianspiering_awesome-dl4nlp_db3cb79f.png","brianspiering","Brian Spiering","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Fbrianspiering_5e274fd4.jpg","Senior Engineer and Educator available for work",null,"San Francisco, CA, USA","bspiering@gmail.com","https:\u002F\u002Fgithub.com\u002Fbrianspiering",1303,255,"2026-03-25T19:19:27",1,"","未说明",{"notes":90,"python":88,"dependencies":91},"该仓库（awesome-dl4nlp）是一个 curated list（资源列表），主要收集了关于深度学习在自然语言处理领域的课程、书籍、教程、框架、论文和数据集链接，本身不是一个可执行的软件工具或代码库，因此 README 中未包含具体的运行环境需求（如操作系统、GPU、内存、Python 版本或依赖库）。用户需根据列表中提到的具体框架（如 TensorFlow, PyTorch, SpaCy 等）或课程代码去查阅相应项目的独立文档以获取环境配置信息。",[],[15],"2026-03-27T02:49:30.150509","2026-04-06T05:44:22.454087",[],[]]