[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-mikeroyal--Machine-Learning-Guide":3,"tool-mikeroyal--Machine-Learning-Guide":61},[4,18,26,36,44,53],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":17},4358,"openclaw","openclaw\u002Fopenclaw","OpenClaw 是一款专为个人打造的本地化 AI 助手，旨在让你在自己的设备上拥有完全可控的智能伙伴。它打破了传统 AI 助手局限于特定网页或应用的束缚，能够直接接入你日常使用的各类通讯渠道，包括微信、WhatsApp、Telegram、Discord、iMessage 等数十种平台。无论你在哪个聊天软件中发送消息，OpenClaw 都能即时响应，甚至支持在 macOS、iOS 和 Android 设备上进行语音交互，并提供实时的画布渲染功能供你操控。\n\n这款工具主要解决了用户对数据隐私、响应速度以及“始终在线”体验的需求。通过将 AI 部署在本地，用户无需依赖云端服务即可享受快速、私密的智能辅助，真正实现了“你的数据，你做主”。其独特的技术亮点在于强大的网关架构，将控制平面与核心助手分离，确保跨平台通信的流畅性与扩展性。\n\nOpenClaw 非常适合希望构建个性化工作流的技术爱好者、开发者，以及注重隐私保护且不愿被单一生态绑定的普通用户。只要具备基础的终端操作能力（支持 macOS、Linux 及 Windows WSL2），即可通过简单的命令行引导完成部署。如果你渴望拥有一个懂你",349277,3,"2026-04-06T06:32:30",[13,14,15,16],"Agent","开发框架","图像","数据工具","ready",{"id":19,"name":20,"github_repo":21,"description_zh":22,"stars":23,"difficulty_score":10,"last_commit_at":24,"category_tags":25,"status":17},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,"2026-04-05T11:01:52",[14,15,13],{"id":27,"name":28,"github_repo":29,"description_zh":30,"stars":31,"difficulty_score":32,"last_commit_at":33,"category_tags":34,"status":17},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",152630,2,"2026-04-12T23:33:54",[14,13,35],"语言模型",{"id":37,"name":38,"github_repo":39,"description_zh":40,"stars":41,"difficulty_score":32,"last_commit_at":42,"category_tags":43,"status":17},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",108322,"2026-04-10T11:39:34",[14,15,13],{"id":45,"name":46,"github_repo":47,"description_zh":48,"stars":49,"difficulty_score":32,"last_commit_at":50,"category_tags":51,"status":17},6121,"gemini-cli","google-gemini\u002Fgemini-cli","gemini-cli 是一款由谷歌推出的开源 AI 命令行工具，它将强大的 Gemini 大模型能力直接集成到用户的终端环境中。对于习惯在命令行工作的开发者而言，它提供了一条从输入提示词到获取模型响应的最短路径，无需切换窗口即可享受智能辅助。\n\n这款工具主要解决了开发过程中频繁上下文切换的痛点，让用户能在熟悉的终端界面内直接完成代码理解、生成、调试以及自动化运维任务。无论是查询大型代码库、根据草图生成应用，还是执行复杂的 Git 操作，gemini-cli 都能通过自然语言指令高效处理。\n\n它特别适合广大软件工程师、DevOps 人员及技术研究人员使用。其核心亮点包括支持高达 100 万 token 的超长上下文窗口，具备出色的逻辑推理能力；内置 Google 搜索、文件操作及 Shell 命令执行等实用工具；更独特的是，它支持 MCP（模型上下文协议），允许用户灵活扩展自定义集成，连接如图像生成等外部能力。此外，个人谷歌账号即可享受免费的额度支持，且项目基于 Apache 2.0 协议完全开源，是提升终端工作效率的理想助手。",100752,"2026-04-10T01:20:03",[52,13,15,14],"插件",{"id":54,"name":55,"github_repo":56,"description_zh":57,"stars":58,"difficulty_score":32,"last_commit_at":59,"category_tags":60,"status":17},4721,"markitdown","microsoft\u002Fmarkitdown","MarkItDown 是一款由微软 AutoGen 团队打造的轻量级 Python 工具，专为将各类文件高效转换为 Markdown 格式而设计。它支持 PDF、Word、Excel、PPT、图片（含 OCR）、音频（含语音转录）、HTML 乃至 YouTube 链接等多种格式的解析，能够精准提取文档中的标题、列表、表格和链接等关键结构信息。\n\n在人工智能应用日益普及的今天，大语言模型（LLM）虽擅长处理文本，却难以直接读取复杂的二进制办公文档。MarkItDown 恰好解决了这一痛点，它将非结构化或半结构化的文件转化为模型“原生理解”且 Token 效率极高的 Markdown 格式，成为连接本地文件与 AI 分析 pipeline 的理想桥梁。此外，它还提供了 MCP（模型上下文协议）服务器，可无缝集成到 Claude Desktop 等 LLM 应用中。\n\n这款工具特别适合开发者、数据科学家及 AI 研究人员使用，尤其是那些需要构建文档检索增强生成（RAG）系统、进行批量文本分析或希望让 AI 助手直接“阅读”本地文件的用户。虽然生成的内容也具备一定可读性，但其核心优势在于为机器",93400,"2026-04-06T19:52:38",[52,14],{"id":62,"github_repo":63,"name":64,"description_en":65,"description_zh":66,"ai_summary_zh":66,"readme_en":67,"readme_zh":68,"quickstart_zh":69,"use_case_zh":70,"hero_image_url":71,"owner_login":72,"owner_name":73,"owner_avatar_url":74,"owner_bio":75,"owner_company":76,"owner_location":77,"owner_email":76,"owner_twitter":78,"owner_website":79,"owner_url":80,"languages":81,"stars":86,"forks":87,"last_commit_at":88,"license":76,"difficulty_score":89,"env_os":90,"env_gpu":91,"env_ram":91,"env_deps":92,"category_tags":95,"github_topics":96,"view_count":32,"oss_zip_url":76,"oss_zip_packed_at":76,"status":17,"created_at":117,"updated_at":118,"faqs":119,"releases":120},7052,"mikeroyal\u002FMachine-Learning-Guide","Machine-Learning-Guide","Machine learning Guide. Learn all about Machine Learning Tools, Libraries, Frameworks, Large Language Models (LLMs), and Training Models.","Machine-Learning-Guide 是一份全面且持续更新的机器学习学习指南，旨在帮助开发者系统掌握从基础理论到前沿应用的各类知识。它解决了机器学习领域技术栈繁杂、学习资源分散的痛点，将庞大的生态系统整理得井井有条。内容涵盖主流开发框架（如 PyTorch、TensorFlow）、大语言模型（LLMs）的训练与本地部署工具、核心算法详解，以及计算机视觉、自然语言处理、强化学习和生物信息学等垂直领域的开发指引。\n\n除了技术工具，该指南还精心汇集了课程认证、专业书籍、YouTube 教程及多语言（Python、C++、Julia 等）开发资源，甚至提供了将文档转换为 PDF 的实用技巧。无论是刚入门的学生、希望提升效率的软件工程师，还是从事前沿探索的研究人员，都能在这里找到适合自己的学习路径和开发利器。其独特的亮点在于对大模型生态的及时跟进，以及对跨语言、跨平台开发环境的细致梳理，是通往高效机器学习开发之路的可靠地图。","\u003Ch1 align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_fa28eb9b69ce.png\">\n  \u003Cbr \u002F>\n  Machine Learning Guide\n\u003C\u002Fh1>\n\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fmikeroyal?tab=followers\">\n         \u003Cimg alt=\"followers\" title=\"Follow me for Updates\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_e273f3008dc5.png\"\u002F>\u003C\u002Fa> \t\n\n![Maintenance](https:\u002F\u002Fimg.shields.io\u002Fmaintenance\u002Fyes\u002F2024?style=for-the-badge)\n![Last-Commit](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flast-commit\u002Fmikeroyal\u002Fmachine-learning-guide?style=for-the-badge)\n\n#### A guide covering Machine Learning including the applications, libraries and tools that will make you better and more efficient with Machine Learning development.\n\n **Note: You can easily convert this markdown file to a PDF in [VSCode](https:\u002F\u002Fcode.visualstudio.com\u002F) using this handy extension [Markdown PDF](https:\u002F\u002Fmarketplace.visualstudio.com\u002Fitems?itemName=yzane.markdown-pdf).**\n \n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_955586009a84.png\">\n\n**Machine Learning\u002FDeep Learning Frameworks.**\n\n# Table of Contents\n\n1. [Learning Resources for ML](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#learning-resources-for-ML)\n\n     - [Developer Resources](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#developer-resources)\n     - [Courses & Certifications](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#courses--certifications)\n     - [Books](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#books)\n     - [YouTube Tutorials](#youtube-tutorials)\n\n2. [ML Frameworks, Libraries, and Tools](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#ML-frameworks-libraries-and-tools)\n\n    - [LLMs Training Frameworks](#llm-training-frameworks)\n    - [Tools for deploying LLMs](#tools-for-deploying-llm)\n    - [Running Large Language Models (LLMs) Locally](#running-llms-locally)\n    \n3. [Algorithms](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#Algorithms)\n\n4. [PyTorch Development](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#pytorch-development)\n\n5. [TensorFlow Development](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#tensorflow-development)\n\n6. [Core ML Development](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#core-ml-development)\n\n7. [Deep Learning Development](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#Deep-Learning-Development)\n\n8. [Reinforcement Learning Development](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#Reinforcement-Learning-Development)\n\n9. [Computer Vision Development](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#computer-vision-development)\n\n10. [Natural Language Processing (NLP) Development](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#nlp-development)\n\n11. [Bioinformatics](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#bioinformatics)\n\n12. [CUDA Development](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#cuda-development)\n\n13. [MATLAB Development](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#matlab-development)\n\n14. [C\u002FC++ Development](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#cc-development)\n\n15. [Java Development](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#java-development)\n\n16. [Python Development](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#python-development)\n\n17. [Scala Development](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#scala-development)\n\n18. [R Development](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#r-development)\n\n19. [Julia Development](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#julia-development)\n\n\n# Learning Resources for ML\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n[Machine Learning](https:\u002F\u002Fwww.ibm.com\u002Fcloud\u002Flearn\u002Fmachine-learning) is a branch of artificial intelligence (AI) focused on building apps using algorithms that learn from data models and improve their accuracy over time without needing to be programmed.\n\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_95dd8ee4e1dd.jpeg\">\n\n### Developer Resources\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n- [Natural Language Processing (NLP) Best Practices by Microsoft](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fnlp-recipes)\n\n- [The Autonomous Driving Cookbook by Microsoft](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002FAutonomousDrivingCookbook)\n\n- [Azure Machine Learning - ML as a Service | Microsoft Azure](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fmachine-learning\u002F)\n\n- [How to run Jupyter Notebooks in your Azure Machine Learning workspace](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fmachine-learning\u002Fhow-to-run-jupyter-notebooks)\n\n- [Machine Learning and Artificial Intelligence | Amazon Web Services](https:\u002F\u002Faws.amazon.com\u002Fmachine-learning\u002F)\n\n- [Scheduling Jupyter notebooks on Amazon SageMaker ephemeral instances](https:\u002F\u002Faws.amazon.com\u002Fblogs\u002Fmachine-learning\u002Fscheduling-jupyter-notebooks-on-sagemaker-ephemeral-instances\u002F)\n\n- [AI & Machine Learning | Google Cloud](https:\u002F\u002Fcloud.google.com\u002Fproducts\u002Fai\u002F)\n\n- [Using Jupyter Notebooks with Apache Spark on Google Cloud](https:\u002F\u002Fcloud.google.com\u002Fblog\u002Fproducts\u002Fgcp\u002Fgoogle-cloud-platform-for-data-scientists-using-jupyter-notebooks-with-apache-spark-on-google-cloud)\n\n- [Machine Learning | Apple Developer](https:\u002F\u002Fdeveloper.apple.com\u002Fmachine-learning\u002F)\n\n- [Artificial Intelligence & Autopilot | Tesla](https:\u002F\u002Fwww.tesla.com\u002FAI)\n\n- [Meta AI Tools | Facebook](https:\u002F\u002Fai.facebook.com\u002Ftools\u002F)\n\n- [PyTorch Tutorials](https:\u002F\u002Fpytorch.org\u002Ftutorials\u002F)\n\n- [TensorFlow Tutorials](https:\u002F\u002Fwww.tensorflow.org\u002Ftutorials)\n\n- [JupyterLab](https:\u002F\u002Fjupyterlab.readthedocs.io\u002F)\n\n- [Stable Diffusion with Core ML on Apple Silicon](https:\u002F\u002Fmachinelearning.apple.com\u002Fresearch\u002Fstable-diffusion-coreml-apple-silicon)\n\n### Courses & Certifications\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n- [Machine Learning by Stanford University by Andrew Ng | Coursera](https:\u002F\u002Fwww.coursera.org\u002Flearn\u002Fmachine-learning)\n\n- [AWS Training and Certification for Machine Learning (ML) Courses](https:\u002F\u002Faws.amazon.com\u002Ftraining\u002Flearning-paths\u002Fmachine-learning\u002F)\n\n- [Machine Learning Scholarship Program for Microsoft Azure | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fscholarships\u002Fmachine-learning-scholarship-microsoft-azure)\n\n- [Microsoft Certified: Azure Data Scientist Associate](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Flearn\u002Fcertifications\u002Fazure-data-scientist)\n\n- [Microsoft Certified: Azure AI Engineer Associate](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Flearn\u002Fcertifications\u002Fazure-ai-engineer)\n\n- [Azure Machine Learning training and deployment](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fdevops\u002Fpipelines\u002Ftargets\u002Fazure-machine-learning)\n\n- [Learning Machine learning and artificial intelligence from Google Cloud Training](https:\u002F\u002Fcloud.google.com\u002Ftraining\u002Fmachinelearning-ai)\n\n- [Machine Learning Crash Course for Google Cloud](https:\u002F\u002Fdevelopers.google.com\u002Fmachine-learning\u002Fcrash-course\u002F)\n\n- [Machine Learning Courses Online | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fmachine-learning\u002F)\n\n- [Machine Learning Courses Online | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=machine%20learning&)\n\n- [Learn Machine Learning with Online Courses and Classes | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fmachine-learning)\n\n### Books\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n* [Introduction To Machine Learning (PDF)](https:\u002F\u002Fai.stanford.edu\u002F~nilsson\u002FMLBOOK.pdf)\n \n* [Artificial Intelligence: A Modern Approach by Stuart J. Russel and Peter Norvig](https:\u002F\u002Fwww.amazon.com\u002FArtificial-Intelligence-A-Modern-Approach\u002Fdp\u002F0134610997\u002Fref=sr_1_1?dchild=1&keywords=artificial+intelligence+a+modern+approach&qid=1626728093&sr=8-1)\n \n* [Deep Learning by Ian Goodfellow, Yoshoua Bengio, and Aaron Courville](https:\u002F\u002Fwww.deeplearningbook.org\u002F)\n \n* [The Hundred-Page Machine Learning Book by Andriy Burkov](https:\u002F\u002Fthemlbook.com\u002Fwiki\u002Fdoku.php)\n \n    - [Hundred-Page Machine Learning Book on GitHub](https:\u002F\u002Fgithub.com\u002Faburkov\u002FtheMLbook)\n \n* [Machine Learning by Tom M. Mitchell](https:\u002F\u002Fwww.cs.cmu.edu\u002F~tom\u002FNewChapters.html)\n \n* [Programming Collective Intelligence: Building Smart Web 2.0 Applications by Toby Segaran](https:\u002F\u002Fwww.amazon.com\u002FProgramming-Collective-Intelligence-Building-Applications\u002Fdp\u002F0596529325\u002Fref=sr_1_1?crid=8EI42XMXESGB&keywords=Programming+Collective+Intelligence%3A+Building+Smart+Web+2.0+Applications&qid=1654318595&sprefix=programming+collective+intelligence+building+smart+web+2.0+applications%2Caps%2C194&sr=8-1)\n \n* [Machine Learning: An Algorithmic Perspective, Second Edition](https:\u002F\u002Fwww.amazon.com\u002FMachine-Learning-Algorithmic-Perspective-Recognition\u002Fdp\u002F1466583282\u002Fref=sr_1_8?crid=2RIQ8OMMASS3&keywords=Pattern+Recognition+and+Machine+Learning&qid=1654318681&sprefix=pattern+recognition+and+machine+learning%2Caps%2C184&sr=8-8)\n\n* [Pattern Recognition and Machine Learning by Christopher M. Bishop](https:\u002F\u002Fwww.amazon.com\u002FPattern-Recognition-Learning-Information-Statistics\u002Fdp\u002F1493938436\u002Fref=sr_1_4?crid=2RIQ8OMMASS3&keywords=Pattern+Recognition+and+Machine+Learning&qid=1654318681&sprefix=pattern+recognition+and+machine+learning%2Caps%2C184&sr=8-4)\n \n * [Natural Language Processing with Python by Steven Bird, Ewan Klein, and Edward Loper](https:\u002F\u002Fwww.amazon.com\u002FNatural-Language-Processing-Python-Analyzing\u002Fdp\u002F0596516495\u002Fref=sr_1_1?crid=O4XSCF3CNIBN&keywords=Natural+Language+Processing+with+Python&qid=1654318757&sprefix=natural+language+processing+with+python%2Caps%2C285&sr=8-1)\n \n * [Python Machine Learning: A Technical Approach to Machine Learning for Beginners by Leonard Eddison](https:\u002F\u002Fwww.amazon.com\u002FPython-Machine-Learning-Technical-Beginners\u002Fdp\u002F1986340872\u002Fref=sr_1_1?crid=1W5X2WV05GDQK&keywords=Python+Machine+Learning%3A+A+Technical+Approach+to+Machine+Learning+for+Beginners&qid=1654318782&sprefix=python+machine+learning+a+technical+approach+to+machine+learning+for+beginners%2Caps%2C212&sr=8-1)\n \n * [Bayesian Reasoning and Machine Learning by David Barber](https:\u002F\u002Fwww.amazon.com\u002FBayesian-Reasoning-Machine-Learning-Barber\u002Fdp\u002F0521518148\u002Fref=sr_1_1?crid=1J054T5MUCD20&keywords=Bayesian+Reasoning+and+Machine+Learning&qid=1654318807&sprefix=bayesian+reasoning+and+machine+learning%2Caps%2C179&sr=8-1)\n  \n * [Machine Learning for Absolute Beginners: A Plain English Introduction by Oliver Theobald](https:\u002F\u002Fwww.amazon.com\u002FMachine-Learning-Absolute-Beginners-Introduction-ebook\u002Fdp\u002FB08RWBSKQB\u002Fref=sr_1_1?crid=1JBS4KEHTY6I5&keywords=Machine+Learning+for+Absolute+Beginners%3A+A+Plain+English+Introduction&qid=1654318861&sprefix=machine+learning+for+absolute+beginners+a+plain+english+introduction%2Caps%2C168&sr=8-1)\n   \n * [Machine Learning in Action by Ben Wilson](https:\u002F\u002Fwww.amazon.com\u002FMachine-Learning-Engineering-Action-Wilson\u002Fdp\u002F1617298719\u002Fref=sr_1_1?crid=6S9F2MJHAQX1&keywords=Machine+Learning+in+Action&qid=1654318897&sprefix=machine+learning+in+action%2Caps%2C174&sr=8-1)\n    \n * [Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems by Aurélien Géron](https:\u002F\u002Fwww.amazon.com\u002FHands-Machine-Learning-Scikit-Learn-TensorFlow\u002Fdp\u002F1492032646\u002Fref=sr_1_6?crid=2RIQ8OMMASS3&keywords=Pattern+Recognition+and+Machine+Learning&qid=1654318681&sprefix=pattern+recognition+and+machine+learning%2Caps%2C184&sr=8-6)\n     \n * [Introduction to Machine Learning with Python: A Guide for Data Scientists by Andreas C. Müller & Sarah Guido](https:\u002F\u002Fwww.amazon.com\u002FIntroduction-Machine-Learning-Python-Scientists\u002Fdp\u002F1449369413\u002Fref=sr_1_1?crid=3SGFHBBU06GB6&keywords=Introduction+to+Machine+Learning+with+Python%3A+A+Guide+for+Data+Scientists&qid=1654318969&sprefix=introduction+to+machine+learning+with+python+a+guide+for+data+scientists%2Caps%2C181&sr=8-1)\n \n * [Machine Learning for Hackers: Case Studies and Algorithms to Get you Started by Drew Conway and John Myles White](https:\u002F\u002Fwww.amazon.com\u002FMachine-Learning-Hackers-Studies-Algorithms\u002Fdp\u002F1449303714\u002Fref=sr_1_1?crid=2PQABQ4T9B8K5&keywords=Machine+Learning+for+Hackers%3A+Case+Studies+and+Algorithms+to+Get+you+Started&qid=1654318629&sprefix=machine+learning+for+hackers+case+studies+and+algorithms+to+get+you+started%2Caps%2C162&sr=8-1)\n \n * [The Elements of Statistical Learning: Data Mining, Inference, and Prediction by Trevor Hastie, Robert Tibshirani, and Jerome Friedman](https:\u002F\u002Fwww.amazon.com\u002FElements-Statistical-Learning-Prediction-Statistics\u002Fdp\u002F0387848576\u002Fref=sr_1_1?crid=1HOK9M9GFHTK9&keywords=The+Elements+of+Statistical+Learning%3A+Data+Mining%2C+Inference%2C+and+Prediction&qid=1654318661&sprefix=the+elements+of+statistical+learning+data+mining%2C+inference%2C+and+prediction+%2Caps%2C215&sr=8-1)\n\n* [Distributed Machine Learning Patterns](https:\u002F\u002Fgithub.com\u002Fterrytangyuan\u002Fdistributed-ml-patterns)  - Book (free to read online) + Code\n* [Real World Machine Learning](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Freal-world-machine-learning) [Free Chapters]\n* [An Introduction To Statistical Learning](https:\u002F\u002Fwww-bcf.usc.edu\u002F~gareth\u002FISL\u002F) - Book + R Code\n* [Elements of Statistical Learning](https:\u002F\u002Fweb.stanford.edu\u002F~hastie\u002FElemStatLearn\u002F) - Book\n* [Think Bayes](https:\u002F\u002Fgreenteapress.com\u002Fwp\u002Fthink-bayes\u002F) - Book + Python Code\n* [Mining Massive Datasets](https:\u002F\u002Finfolab.stanford.edu\u002F~ullman\u002Fmmds\u002Fbook.pdf)\n* [A First Encounter with Machine Learning](https:\u002F\u002Fwww.ics.uci.edu\u002F~welling\u002Fteaching\u002F273ASpring10\u002FIntroMLBook.pdf)\n* [Introduction to Machine Learning](https:\u002F\u002Falex.smola.org\u002Fdrafts\u002Fthebook.pdf) - Alex Smola and S.V.N. Vishwanathan\n* [A Probabilistic Theory of Pattern Recognition](https:\u002F\u002Fwww.szit.bme.hu\u002F~gyorfi\u002Fpbook.pdf)\n* [Introduction to Information Retrieval](https:\u002F\u002Fnlp.stanford.edu\u002FIR-book\u002Fpdf\u002Firbookprint.pdf)\n* [Forecasting: principles and practice](https:\u002F\u002Fotexts.com\u002Ffpp2\u002F)\n* [Introduction to Machine Learning](https:\u002F\u002Farxiv.org\u002Fpdf\u002F0904.3664v1.pdf) - Amnon Shashua\n* [Reinforcement Learning](https:\u002F\u002Fwww.intechopen.com\u002Fbooks\u002Freinforcement_learning)\n* [Machine Learning](https:\u002F\u002Fwww.intechopen.com\u002Fbooks\u002Fmachine_learning)\n* [A Quest for AI](https:\u002F\u002Fai.stanford.edu\u002F~nilsson\u002FQAI\u002Fqai.pdf)\n* [R Programming for Data Science](https:\u002F\u002Fleanpub.com\u002Frprogramming)\n* [Data Mining - Practical Machine Learning Tools and Techniques](https:\u002F\u002Fcdn.preterhuman.net\u002Ftexts\u002Fscience_and_technology\u002Fartificial_intelligence\u002FData%20Mining%20Practical%20Machine%20Learning%20Tools%20and%20Techniques%202d%20ed%20-%20Morgan%20Kaufmann.pdf) \n* [Machine Learning with TensorFlow](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fmachine-learning-with-tensorflow) \n* [Machine Learning Systems](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fmachine-learning-systems) \n* [Foundations of Machine Learning](https:\u002F\u002Fcs.nyu.edu\u002F~mohri\u002Fmlbook\u002F) - Mehryar Mohri, Afshin Rostamizadeh, and Ameet Talwalkar\n* [AI-Powered Search](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fai-powered-search) - Trey Grainger, Doug Turnbull, Max Irwin - \n* [Ensemble Methods for Machine Learning](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fensemble-methods-for-machine-learning) - Gautam Kunapuli \n* [Machine Learning Engineering in Action](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fmachine-learning-engineering-in-action) - Ben Wilson \n* [Privacy-Preserving Machine Learning](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fprivacy-preserving-machine-learning) - J. Morris Chang, Di Zhuang, G. Dumindu Samaraweera \n* [Automated Machine Learning in Action](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fautomated-machine-learning-in-action) - Qingquan Song, Haifeng Jin, and Xia Hu \n* [Distributed Machine Learning Patterns](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fdistributed-machine-learning-patterns) - Yuan Tang \n* [Managing Machine Learning Projects: From design to deployment](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fmanaging-machine-learning-projects) - Simon Thompson\n* [Causal Machine Learning](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fcausal-machine-learning) - Robert Ness \n* [Bayesian Optimization in Action](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fbayesian-optimization-in-action) - Quan Nguyen \n* [Machine Learning Algorithms in Depth](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fmachine-learning-algorithms-in-depth)) - Vadim Smolyakov \n* [Optimization Algorithms](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Foptimization-algorithms) - Alaa Khamis\n* [Practical Gradient Boosting](https:\u002F\u002Fwww.amazon.com\u002Fdp\u002FB0BL1HRD6Z) by Guillaume Saupin\n\n### YouTube Tutorials\n\n[Back to the Top](#table-of-contents)\n\n[![Andrew Ng: Opportunities in AI - Standford 2023](https:\u002F\u002Fytcards.demolab.com\u002F?id=5p248yoa3oE&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"Andrew Ng: Opportunities in AI - Standford 2023\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=5p248yoa3oE)\n[![How Does AI Actually Work?](https:\u002F\u002Fytcards.demolab.com\u002F?id=3ihjz7g1OQM&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"How Does AI Actually Work?\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=3ihjz7g1OQM)\n[![AI vs Machine Learning](https:\u002F\u002Fytcards.demolab.com\u002F?id=4RixMPF4xis&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"AI vs Machine Learning\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=4RixMPF4xis)\n[![Machine Learning vs Deep Learning](https:\u002F\u002Fytcards.demolab.com\u002F?id=q6kJ71tEYqM&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"Machine Learning vs Deep Learning\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=q6kJ71tEYqM)\n[![What are Transformers (Machine Learning Model)?](https:\u002F\u002Fytcards.demolab.com\u002F?id=ZXiruGOCn9s&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"What are Transformers (Machine Learning Model)?\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=ZXiruGOCn9s)\n[![But what is a neural network? | Chapter 1, Deep learning](https:\u002F\u002Fytcards.demolab.com\u002F?id=aircAruvnKk&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"But what is a neural network? | Chapter 1, Deep learning\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=aircAruvnKk)\n[![Advice for machine learning beginners | Andrej Karpathy and Lex Fridman](https:\u002F\u002Fytcards.demolab.com\u002F?id=I2ZK3ngNvvI&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"Advice for machine learning beginners | Andrej Karpathy and Lex Fridman\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=I2ZK3ngNvvI)\n[![Machine Learning Explained in 100 Seconds](https:\u002F\u002Fytcards.demolab.com\u002F?id=PeMlggyqz0Y&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"Machine Learning Explained in 100 Seconds\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=PeMlggyqz0Y)\n[![How to learn AI and ML in 2023 - A complete roadmap](https:\u002F\u002Fytcards.demolab.com\u002F?id=KEB-w9DUdCw&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"How to learn AI and ML in 2023 - A complete roadmap\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=KEB-w9DUdCw)\n[![PyTorch for Deep Learning & Machine Learning – Full Course](https:\u002F\u002Fytcards.demolab.com\u002F?id=V_xro1bcAuA&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"PyTorch for Deep Learning & Machine Learning – Full Course\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=V_xro1bcAuA)\n[![Deep Learning for Computer Vision with Python and TensorFlow](https:\u002F\u002Fytcards.demolab.com\u002F?id=IA3WxTTPXqQ&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"Deep Learning for Computer Vision with Python and TensorFlow\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=IA3WxTTPXqQ)\n[![How Large Language Models Work](https:\u002F\u002Fytcards.demolab.com\u002F?id=5sLYAQS9sWQ&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"How Large Language Models Work\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=5sLYAQS9sWQ)\n[![What are Large Language Models (LLMs)?](https:\u002F\u002Fytcards.demolab.com\u002F?id=iR2O2GPbB0E&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"What are Large Language Models (LLMs)?\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=iR2O2GPbB0E)\n[![Introduction to large language model](https:\u002F\u002Fytcards.demolab.com\u002F?id=zizonToFXDs&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"Introduction to large language model\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=zizonToFXDs)\n[![Create a Large Language Model from Scratch with Python](https:\u002F\u002Fytcards.demolab.com\u002F?id=UU1WVnMk4E8&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"Create a Large Language Model from Scratch with Python\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=UU1WVnMk4E8)\n\n# ML Frameworks, Libraries, and Tools\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n[TensorFlow](https:\u002F\u002Fwww.tensorflow.org) is an end-to-end open source platform for machine learning. It has a comprehensive, flexible ecosystem of tools, libraries and community resources that lets researchers push the state-of-the-art in ML and developers easily build and deploy ML powered applications.\n\n[Keras](https:\u002F\u002Fkeras.io) is a high-level neural networks API, written in Python and capable of running on top of TensorFlow, CNTK, or Theano.It was developed with a focus on enabling fast experimentation. It is capable of running on top of TensorFlow, Microsoft Cognitive Toolkit, R, Theano, or PlaidML.\n\n[PyTorch](https:\u002F\u002Fpytorch.org) is a library for deep learning on irregular input data such as graphs, point clouds, and manifolds. Primarily developed by Facebook's AI Research lab.\n\n[Amazon SageMaker](https:\u002F\u002Faws.amazon.com\u002Fsagemaker\u002F) is a fully managed service that provides every developer and data scientist with the ability to build, train, and deploy machine learning (ML) models quickly. SageMaker removes the heavy lifting from each step of the machine learning process to make it easier to develop high quality models.\n\n[Azure Databricks](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fdatabricks\u002F) is a fast and collaborative Apache Spark-based big data analytics service designed for data science and data engineering. Azure Databricks, sets up your Apache Spark environment in minutes, autoscale, and collaborate on shared projects in an interactive workspace. Azure Databricks supports Python, Scala, R, Java, and SQL, as well as data science frameworks and libraries including TensorFlow, PyTorch, and scikit-learn.\n\n[Microsoft Cognitive Toolkit (CNTK)](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fcognitive-toolkit\u002F) is an open-source toolkit for commercial-grade distributed deep learning. It describes neural networks as a series of computational steps via a directed graph. CNTK allows the user to easily realize and combine popular model types such as feed-forward DNNs, convolutional neural networks (CNNs) and recurrent neural networks (RNNs\u002FLSTMs). CNTK implements stochastic gradient descent (SGD, error backpropagation) learning with automatic differentiation and parallelization across multiple GPUs and servers.\n\n[Apple CoreML](https:\u002F\u002Fdeveloper.apple.com\u002Fdocumentation\u002Fcoreml) is a framework that helps integrate machine learning models into your app. Core ML provides a unified representation for all models. Your app uses Core ML APIs and user data to make predictions, and to train or fine-tune models, all on the user's device. A model is the result of applying a machine learning algorithm to a set of training data. You use a model to make predictions based on new input data.\n\n[Apache OpenNLP](https:\u002F\u002Fopennlp.apache.org\u002F) is an open-source library for a machine learning based toolkit used in the processing of natural language text. It features an API for use cases like [Named Entity Recognition](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FNamed-entity_recognition), [Sentence Detection](), [POS(Part-Of-Speech) tagging](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FPart-of-speech_tagging), [Tokenization](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FTokenization_(data_security)) [Feature extraction](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FFeature_extraction), [Chunking](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FChunking_(psychology)), [Parsing](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FParsing), and [Coreference resolution](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FCoreference).\n\n[Apache Airflow](https:\u002F\u002Fairflow.apache.org) is an open-source workflow management platform created by the community to programmatically author, schedule and monitor workflows. Install. Principles. Scalable. Airflow has a modular architecture and uses a message queue to orchestrate an arbitrary number of workers. Airflow is ready to scale to infinity.\n\n[Open Neural Network Exchange(ONNX)](https:\u002F\u002Fgithub.com\u002Fonnx) is an open ecosystem that empowers AI developers to choose the right tools as their project evolves. ONNX provides an open source format for AI models, both deep learning and traditional ML. It defines an extensible computation graph model, as well as definitions of built-in operators and standard data types.\n\n[Apache MXNet](https:\u002F\u002Fmxnet.apache.org\u002F) is a deep learning framework designed for both efficiency and flexibility. It allows you to mix symbolic and imperative programming to maximize efficiency and productivity. At its core, MXNet contains a dynamic dependency scheduler that automatically parallelizes both symbolic and imperative operations on the fly. A graph optimization layer on top of that makes symbolic execution fast and memory efficient. MXNet is portable and lightweight, scaling effectively to multiple GPUs and multiple machines. Support for Python, R, Julia, Scala, Go, Javascript and more.\n\n[AutoGluon](https:\u002F\u002Fautogluon.mxnet.io\u002Findex.html) is toolkit for Deep learning that automates machine learning tasks enabling you to easily achieve strong predictive performance in your applications. With just a few lines of code, you can train and deploy high-accuracy deep learning models on tabular, image, and text data.\n\n[Anaconda](https:\u002F\u002Fwww.anaconda.com\u002F) is a very popular Data Science platform for machine learning and deep learning that enables users to develop models, train them, and deploy them.\n\n[PlaidML](https:\u002F\u002Fgithub.com\u002Fplaidml\u002Fplaidml) is an advanced and portable tensor compiler for enabling deep learning on laptops, embedded devices, or other devices where the available computing hardware is not well supported or the available software stack contains unpalatable license restrictions.\n\n[OpenCV](https:\u002F\u002Fopencv.org) is a highly optimized library with focus on real-time computer vision applications. The C++, Python, and Java interfaces support Linux, MacOS, Windows, iOS, and Android.\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) is a Python module for machine learning built on top of SciPy, NumPy, and matplotlib, making it easier to apply robust and simple implementations of many popular machine learning algorithms.\n\n[Weka](https:\u002F\u002Fwww.cs.waikato.ac.nz\u002Fml\u002Fweka\u002F) is an open source machine learning software that can be accessed through a graphical user interface, standard terminal applications, or a Java API. It is widely used for teaching, research, and industrial applications, contains a plethora of built-in tools for standard machine learning tasks, and additionally gives transparent access to well-known toolboxes such as scikit-learn, R, and Deeplearning4j. \n\n[Caffe](https:\u002F\u002Fgithub.com\u002FBVLC\u002Fcaffe) is a deep learning framework made with expression, speed, and modularity in mind. It is developed by Berkeley AI Research (BAIR)\u002FThe Berkeley Vision and Learning Center (BVLC) and community contributors.\n\n[Theano](https:\u002F\u002Fgithub.com\u002FTheano\u002FTheano) is a Python library that allows you to define, optimize, and evaluate mathematical expressions involving multi-dimensional arrays efficiently including tight integration with NumPy.\n\n[nGraph](https:\u002F\u002Fgithub.com\u002FNervanaSystems\u002Fngraph) is an open source C++ library, compiler and runtime for Deep Learning. The nGraph Compiler aims to accelerate developing AI workloads using any deep learning framework and deploying to a variety of hardware targets.It provides the freedom, performance, and ease-of-use to AI developers.\n\n[NVIDIA cuDNN](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcudnn) is a GPU-accelerated library of primitives for [deep neural networks](https:\u002F\u002Fdeveloper.nvidia.com\u002Fdeep-learning). cuDNN provides highly tuned implementations for standard routines such as forward and backward convolution, pooling, normalization, and activation layers. cuDNN accelerates widely used deep learning frameworks, including [Caffe2](https:\u002F\u002Fcaffe2.ai\u002F), [Chainer](https:\u002F\u002Fchainer.org\u002F), [Keras](https:\u002F\u002Fkeras.io\u002F), [MATLAB](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fdeep-learning.html), [MxNet](https:\u002F\u002Fmxnet.incubator.apache.org\u002F), [PyTorch](https:\u002F\u002Fpytorch.org\u002F), and [TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002F).\n\n[Huginn](https:\u002F\u002Fgithub.com\u002Fhuginn\u002Fhuginn) is a self-hosted system for building agents that perform automated tasks for you online. It can read the web, watch for events, and take actions on your behalf. Huginn's Agents create and consume events, propagating them along a directed graph. Think of it as a hackable version of IFTTT or Zapier on your own server. \n\n[Netron](https:\u002F\u002Fnetron.app\u002F) is a viewer for neural network, deep learning and machine learning models. It supports ONNX, TensorFlow Lite, Caffe, Keras, Darknet, PaddlePaddle, ncnn, MNN, Core ML, RKNN, MXNet, MindSpore Lite, TNN, Barracuda, Tengine, CNTK, TensorFlow.js, Caffe2 and UFF.\n\n[Dopamine](https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fdopamine) is a research framework for fast prototyping of reinforcement learning algorithms. \n\n[DALI](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FDALI) is a GPU-accelerated library containing highly optimized building blocks and an execution engine for data processing to accelerate deep learning training and inference applications.\n\n[MindSpore Lite](https:\u002F\u002Fgithub.com\u002Fmindspore-ai\u002Fmindspore) is a new open source deep learning training\u002Finference framework that could be used for mobile, edge and cloud scenarios. \n\n[Darknet](https:\u002F\u002Fgithub.com\u002Fpjreddie\u002Fdarknet) is an open source neural network framework written in C and CUDA. It is fast, easy to install, and supports CPU and GPU computation.\n\n[PaddlePaddle](https:\u002F\u002Fgithub.com\u002FPaddlePaddle\u002FPaddle) is an easy-to-use, efficient, flexible and scalable deep learning platform, which is originally developed by Baidu scientists and engineers for the purpose of applying deep learning to many products at Baidu. \n\n[GoogleNotebookLM](https:\u002F\u002Fblog.google\u002Ftechnology\u002Fai\u002Fnotebooklm-google-ai\u002F) is an experimental AI tool using the power of language models paired with your existing content to gain critical insights, faster. Similar to a virtual research assistant that can summarize facts, explain complex ideas, and brainstorm new connections based on the sources you select.\n\n[Unilm](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Funilm) is a large-scale Self-supervised Pre-training Across Tasks, Languages, and Modalities.\n\n[Semantic Kernel (SK)](https:\u002F\u002Faka.ms\u002Fsemantic-kernel) is a lightweight SDK enabling integration of AI Large Language Models (LLMs) with conventional programming languages. The SK extensible programming model combines natural language semantic functions, traditional code native functions, and embeddings-based memory unlocking new potential and adding value to applications with AI.\n\n[Pandas AI](https:\u002F\u002Fgithub.com\u002Fgventuri\u002Fpandas-ai) is a Python library that integrates generative artificial intelligence capabilities into Pandas, making dataframes conversational.\n\n[NCNN](https:\u002F\u002Fgithub.com\u002FTencent\u002Fncnn) is a high-performance neural network inference framework optimized for the mobile platform. \n\n[MNN](https:\u002F\u002Fgithub.com\u002Falibaba\u002FMNN) is a blazing fast, lightweight deep learning framework, battle-tested by business-critical use cases in Alibaba.\n\n[MediaPipe](https:\u002F\u002Fmediapipe.dev\u002F) is an optimized for end-to-end performance on a wide array of platforms. See demos Learn more Complex on-device ML, simplified We've abstracted away the complexities of making on-device ML customizable, production-ready, and accessible across platforms.\n\n[MegEngine](https:\u002F\u002Fgithub.com\u002FMegEngine) is a fast, scalable, and user friendly deep learning framework with 3 key features: Unified framework for both training and inference.\n\n[ML.NET](https:\u002F\u002Fdot.net\u002Fml) is a machine learning library that is designed as an extensible platform so that you can consume other popular ML frameworks (TensorFlow, ONNX, Infer.NET, and more) and have access to even more machine learning scenarios, like image classification, object detection, and more. \n\n[Ludwig](https:\u002F\u002Fludwig.ai\u002F) is a [declarative machine learning framework](https:\u002F\u002Fludwig-ai.github.io\u002Fludwig-docs\u002Flatest\u002Fuser_guide\u002Fwhat_is_ludwig\u002F#why-declarative-machine-learning-systems) that makes it easy to define machine learning pipelines using a simple and flexible data-driven configuration system. \n\n[MMdnn](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002FMMdnn) is a comprehensive and cross-framework tool to convert, visualize and diagnose deep learning (DL) models. The \"MM\" stands for model management, and \"dnn\" is the acronym of deep neural network. Convert models between Caffe, Keras, MXNet, Tensorflow, CNTK, PyTorch Onnx and CoreML. \n\n[Horovod](https:\u002F\u002Fgithub.com\u002Fhorovod\u002Fhorovod) is a distributed deep learning training framework for TensorFlow, Keras, PyTorch, and Apache MXNet.  \n\n[Vaex](https:\u002F\u002Fvaex.io\u002F) is a high performance Python library for lazy Out-of-Core DataFrames (similar to Pandas), to visualize and explore big tabular datasets.\n\n[GluonTS](https:\u002F\u002Fts.gluon.ai\u002F) is a Python package for probabilistic time series modeling, focusing on deep learning based models, based on [PyTorch](https:\u002F\u002Fpytorch.org\u002F) and [MXNet](https:\u002F\u002Fmxnet.apache.org\u002F).\n\n[MindsDB](http:\u002F\u002Fmindsdb.com\u002F) is a ML-SQL Server enables machine learning workflows for the most powerful databases and data warehouses using SQL.\n\n[Jupyter Notebook](https:\u002F\u002Fjupyter.org\u002F) is an open-source web application that allows you to create and share documents that contain live code, equations, visualizations and narrative text. Jupyter is used widely in industries that do data cleaning and transformation, numerical simulation, statistical modeling, data visualization, data science, and machine learning.\n\n[Apache Spark](https:\u002F\u002Fspark.apache.org\u002F) is a unified analytics engine for large-scale data processing. It provides high-level APIs in Scala, Java, Python, and R, and an optimized engine that supports general computation graphs for data analysis. It also supports a rich set of higher-level tools including Spark SQL for SQL and DataFrames, MLlib for machine learning, GraphX for graph processing, and Structured Streaming for stream processing.\n\n[Apache Spark Connector for SQL Server and Azure SQL](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fsql-spark-connector) is a high-performance connector that enables you to use transactional data in big data analytics and persists results for ad-hoc queries or reporting. The connector allows you to use any SQL database, on-premises or in the cloud, as an input data source or output data sink for Spark jobs.\n\n[Apache PredictionIO](https:\u002F\u002Fpredictionio.apache.org\u002F) is an open source machine learning framework for developers, data scientists, and end users. It supports event collection, deployment of algorithms, evaluation, querying predictive results via REST APIs. It is based on scalable open source services like Hadoop, HBase (and other DBs), Elasticsearch, Spark and implements what is called a Lambda Architecture.\n\n[Cluster Manager for Apache Kafka(CMAK)](https:\u002F\u002Fgithub.com\u002Fyahoo\u002FCMAK) is a tool for managing [Apache Kafka](https:\u002F\u002Fkafka.apache.org\u002F) clusters.\n\n[BigDL](https:\u002F\u002Fbigdl-project.github.io\u002F) is a distributed deep learning library for Apache Spark. With BigDL, users can write their deep learning applications as standard Spark programs, which can directly run on top of existing Spark or Hadoop clusters.\n\n[Eclipse Deeplearning4J (DL4J)](https:\u002F\u002Fdeeplearning4j.konduit.ai\u002F) is a set of projects intended to support all the needs of a JVM-based(Scala, Kotlin, Clojure, and Groovy) deep learning application. This means starting with the raw data, loading and preprocessing it from wherever and whatever format it is in to building and tuning a wide variety of simple and complex deep learning networks.\n\n[Tensorman](https:\u002F\u002Fgithub.com\u002Fpop-os\u002Ftensorman) is a utility for easy management of Tensorflow containers by developed by [System76]( https:\u002F\u002Fsystem76.com).Tensorman allows Tensorflow to operate in an isolated environment that is contained from the rest of the system. This virtual environment can operate independent of the base system, allowing you to use any version of Tensorflow on any version of a Linux distribution that supports the Docker runtime.\n\n[Numba](https:\u002F\u002Fgithub.com\u002Fnumba\u002Fnumba) is an open source, NumPy-aware optimizing compiler for Python sponsored by Anaconda, Inc. It uses the LLVM compiler project to generate machine code from Python syntax. Numba can compile a large subset of numerically-focused Python, including many NumPy functions. Additionally, Numba has support for automatic parallelization of loops, generation of GPU-accelerated code, and creation of ufuncs and C callbacks.\n\n[Chainer](https:\u002F\u002Fchainer.org\u002F) is a Python-based deep learning framework aiming at flexibility. It provides automatic differentiation APIs based on the define-by-run approach (dynamic computational graphs) as well as object-oriented high-level APIs to build and train neural networks. It also supports CUDA\u002FcuDNN using [CuPy](https:\u002F\u002Fgithub.com\u002Fcupy\u002Fcupy) for high performance training and inference.\n\n[XGBoost](https:\u002F\u002Fxgboost.readthedocs.io\u002F) is an optimized distributed gradient boosting library designed to be highly efficient, flexible and portable. It implements machine learning algorithms under the Gradient Boosting framework. XGBoost provides a parallel tree boosting (also known as GBDT, GBM) that solve many data science problems in a fast and accurate way. It supports distributed training on multiple machines, including AWS, GCE, Azure, and Yarn clusters. Also, it can be integrated with Flink, Spark and other cloud dataflow systems.\n\n[cuML](https:\u002F\u002Fgithub.com\u002Frapidsai\u002Fcuml) is a suite of libraries that implement machine learning algorithms and mathematical primitives functions that share compatible APIs with other RAPIDS projects. cuML enables data scientists, researchers, and software engineers to run traditional tabular ML tasks on GPUs without going into the details of CUDA programming. In most cases, cuML's Python API matches the API from scikit-learn.\n\n[Emu](https:\u002F\u002Fcalebwin.github.io\u002Femu) is a GPGPU library for Rust with a focus on portability, modularity, and performance. It's a CUDA-esque compute-specific abstraction over WebGPU providing specific functionality to make WebGPU feel more like CUDA.\n\n[Scalene](https:\u002F\u002Fgithub.com\u002Fplasma-umass\u002Fscalene) is a high-performance CPU, GPU and memory profiler for Python that does a number of things that other Python profilers do not and cannot do. It runs orders of magnitude faster than many other profilers while delivering far more detailed information. \n\n[MLpack](https:\u002F\u002Fmlpack.org\u002F) is a fast, flexible C++ machine learning library written in C++ and built on the [Armadillo](https:\u002F\u002Farma.sourceforge.net\u002F) linear algebra library, the [ensmallen](https:\u002F\u002Fensmallen.org\u002F) numerical optimization library, and parts of [Boost](https:\u002F\u002Fboost.org\u002F). \n\n[Netron](https:\u002F\u002Fnetron.app\u002F) is a viewer for neural network, deep learning and machine learning models. It supports ONNX, TensorFlow Lite, Caffe, Keras, Darknet, PaddlePaddle, ncnn, MNN, Core ML, RKNN, MXNet, MindSpore Lite, TNN, Barracuda, Tengine, CNTK, TensorFlow.js, Caffe2 and UFF.\n\n[Lightning](https:\u002F\u002Fgithub.com\u002FLightning-AI\u002Flightning) is a tool that builds and trains PyTorch models and connect them to the ML lifecycle using Lightning App templates, without handling DIY infrastructure, cost management, scaling, etc..\n\n[OpenNN](https:\u002F\u002Fwww.opennn.net\u002F) is an open-source neural networks library for machine learning. It contains sophisticated algorithms and utilities to deal with many artificial intelligence solutions.\n\n[H20](https:\u002F\u002Fh2o.ai\u002F) is an AI Cloud platform that solves complex business problems and accelerates the discovery of new ideas with results you can understand and trust.\n\n[Gensim](https:\u002F\u002Fgithub.com\u002FRaRe-Technologies\u002Fgensim) is a Python library for topic modelling, document indexing and similarity retrieval with large corpora. Target audience is the natural language processing (NLP) and information retrieval (IR) community.\n\n[llama.cpp](https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fllama.cpp) is a Port of Facebook's LLaMA model in C\u002FC++.\n\n[hmmlearn](https:\u002F\u002Fgithub.com\u002Fhmmlearn\u002Fhmmlearn) is a set of algorithms for unsupervised learning and inference of [Hidden Markov Models](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FHidden_Markov_model). \n\n[Nextjournal](https:\u002F\u002Fnextjournal.com\u002F) is a notebook for reproducible research. It runs anything you can put into a Docker container. Improve your workflow with polyglot notebooks, automatic versioning and real-time collaboration. Save time and money with on-demand provisioning, including GPU support.\n\n[IPython](https:\u002F\u002Fipython.org\u002F) provides a rich architecture for interactive computing with:\n\n  - A powerful interactive shell.\n  - A kernel for [Jupyter](https:\u002F\u002Fjupyter.org\u002F).\n  - Support for interactive data visualization and use of [GUI toolkits](https:\u002F\u002Fipython.org\u002Fipython-doc\u002Fstable\u002Finteractive\u002Freference.html#gui-event-loop-support).\n  - Flexible, [embeddable](https:\u002F\u002Fipython.org\u002Fipython-doc\u002Fstable\u002Finteractive\u002Freference.html#embedding-ipython) interpreters to load into your own projects.\n  - Easy to use, high performance tools for [parallel computing](https:\u002F\u002Fipyparallel.readthedocs.io\u002Fen\u002Flatest\u002F).\n\n[Veles](https:\u002F\u002Fgithub.com\u002FSamsung\u002Fveles) is a Distributed platform for rapid Deep learning application development currently devloped by Samsung.\n\n[DyNet](https:\u002F\u002Fgithub.com\u002Fclab\u002Fdynet) is a neural network library developed by Carnegie Mellon University and many others. It is written in C++ (with bindings in Python) and is designed to be efficient when run on either CPU or GPU, and to work well with networks that have dynamic structures that change for every training instance. These kinds of networks are particularly important in natural language processing tasks, and DyNet has been used to build state-of-the-art systems for syntactic parsing, machine translation, morphological inflection, and many other application areas.\n\n[Ray](https:\u002F\u002Fgithub.com\u002Fray-project\u002Fray) is a unified framework for scaling AI and Python applications. It consists of a core distributed runtime and a toolkit of libraries (Ray AIR) for accelerating ML workloads.\n\n[whisper.cpp](https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fwhisper.cpp) is a high-performance inference of OpenAI's Whisper automatic speech recognition (ASR) model.\n\n[ChatGPT Plus](https:\u002F\u002Fopenai.com\u002Fblog\u002Fchatgpt-plus\u002F) is a pilot subscription plan(**$20\u002Fmonth**) for ChatGPT, a conversational AI that can chat with you, answer follow-up questions, and challenge incorrect assumptions.\n\n[Auto-GPT](https:\u002F\u002Fgithub.com\u002FSignificant-Gravitas\u002FAuto-GPT) is an \"AI agent\" that given a goal in natural language, can attempt to achieve it by breaking it into sub-tasks and using the internet and other tools in an automatic loop. It uses OpenAI's GPT-4 or GPT-3.5 APIs, and is among the first examples of an application using GPT-4 to perform autonomous tasks.\n\n[Chatbot UI by mckaywrigley](https:\u002F\u002Fgithub.com\u002Fmckaywrigley\u002Fchatbot-ui) is an advanced chatbot kit for OpenAI's chat models built on \ntop of Chatbot UI Lite using Next.js, TypeScript, and Tailwind CSS. This version of ChatBot UI supports both GPT-3.5 and GPT-4 models. Conversations are stored locally within your browser. You can export and import conversations to safeguard against data loss. See a [demo](https:\u002F\u002Ftwitter.com\u002Fmckaywrigley\u002Fstatus\u002F1636103188733640704).\n\n[Chatbot UI Lite by mckaywrigley](https:\u002F\u002Fgithub.com\u002Fmckaywrigley\u002Fchatbot-ui-lite) is a simple chatbot starter kit for OpenAI's chat model using Next.js, TypeScript, and Tailwind CSS. See a [demo](https:\u002F\u002Ftwitter.com\u002Fmckaywrigley\u002Fstatus\u002F1636103188733640704).\n\n[MiniGPT-4](https:\u002F\u002Fminigpt-4.github.io\u002F) is an enhancing Vision-language Understanding with Advanced Large Language Models.\n\n[GPT4All](https:\u002F\u002Fgithub.com\u002Fnomic-ai\u002Fgpt4all) is an ecosystem of open-source chatbots trained on a massive collections of clean assistant data including code, stories and dialogue based on [LLaMa](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fllama).\n\n[GPT4All UI](https:\u002F\u002Fgithub.com\u002Fnomic-ai\u002Fgpt4all-ui) is a Flask web application that provides a chat UI for interacting with the GPT4All chatbot. \n\n[Alpaca.cpp](https:\u002F\u002Fgithub.com\u002Fantimatter15\u002Falpaca.cpp) is a fast ChatGPT-like model locally on your device. It combines the [LLaMA foundation model](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fllama) with an [open reproduction](https:\u002F\u002Fgithub.com\u002Ftloen\u002Falpaca-lora) of [Stanford Alpaca](https:\u002F\u002Fgithub.com\u002Ftatsu-lab\u002Fstanford_alpaca) a fine-tuning of the base model to obey instructions (akin to the [RLHF](https:\u002F\u002Fhuggingface.co\u002Fblog\u002Frlhf) used to train ChatGPT) and a set of modifications to [llama.cpp](https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fllama.cpp) to add a chat interface.\n\n[llama.cpp](https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fllama.cpp) is a Port of Facebook's LLaMA model in C\u002FC++.\n\n[OpenPlayground](https:\u002F\u002Fgithub.com\u002Fnat\u002Fopenplayground) is a playfround for running ChatGPT-like models locally on your device.\n\n[Vicuna](https:\u002F\u002Fvicuna.lmsys.org\u002F) is an open source chatbot trained by fine tuning LLaMA. It apparently achieves more than 90% quality of chatgpt and costs $300 to train.\n\n[Yeagar ai](https:\u002F\u002Fgithub.com\u002Fyeagerai\u002Fyeagerai-agent) is a Langchain Agent creator designed to help you build, prototype, and deploy AI-powered agents with ease.\n\n[Vicuna](https:\u002F\u002Fvicuna.lmsys.org\u002F) is created by fine-tuning a LLaMA base model using approximately 70K user-shared conversations gathered from ShareGPT.com with public APIs. To ensure data quality, it convert the HTML back to markdown and filter out some inappropriate or low-quality samples.\n\n[ShareGPT](https:\u002F\u002Fsharegpt.com\u002F) is a place to share your wildest ChatGPT conversations with one click. With 198,404 conversations shared so far. \n\n[FastChat](https:\u002F\u002Fgithub.com\u002Flm-sys\u002FFastChat) is an open platform for training, serving, and evaluating large language model based chatbots.\n\n[Haystack](https:\u002F\u002Fhaystack.deepset.ai\u002F) is an open source NLP framework to interact with your data using Transformer models and LLMs (GPT-4, ChatGPT and alike). It offers production-ready tools to quickly build complex decision making, question answering, semantic search, text generation applications, and more. \n\n[StableLM (Stability AI Language Models)](https:\u002F\u002Fgithub.com\u002FStability-AI\u002FStableLM) is StableLM series of language models and will be continuously updated with new checkpoints. \n\n[Databricks’ Dolly](https:\u002F\u002Fgithub.com\u002Fdatabrickslabs\u002Fdolly) is an instruction-following large language model trained on the Databricks machine learning platform that is licensed for commercial use. \n\n[GPTCach](https:\u002F\u002Fgptcache.readthedocs.io\u002F) is a Library for Creating Semantic Cache for LLM Queries.\n\n[AlaC](https:\u002F\u002Fgithub.com\u002Fgofireflyio\u002Faiac) is an Artificial Intelligence Infrastructure-as-Code Generator. \n\n[Adrenaline](https:\u002F\u002Fuseadrenaline.com\u002F) is a tool that lets you talk to your codebase. It's powered by static analysis, vector search, and large language models.\n\n[OpenAssistant](https:\u002F\u002Fopen-assistant.io\u002F) is a chat-based assistant that understands tasks, can interact with third-party systems, and retrieve information dynamically to do so. \n\n[DoctorGPT](https:\u002F\u002Fgithub.com\u002Fingyamilmolinar\u002Fdoctorgpt) is a lightweight self-contained binary that monitors your application logs for problems and diagnoses them.\n\n[HttpGPT](https:\u002F\u002Fgithub.com\u002Flucoiso\u002FUEHttpGPT\u002Freleases) is an Unreal Engine 5 plugin that facilitates integration with OpenAI's GPT based services (ChatGPT and DALL-E) through asynchronous REST requests, making it easy for developers to communicate with these services. It also includes Editor Tools to integrate Chat GPT and DALL-E image generation directly in the Engine.\n\n[PaLM 2](https:\u002F\u002Fai.google\u002Fdiscover\u002Fpalm2) is a next generation large language model that builds on Google’s legacy of breakthrough research in machine learning and responsible AI. It includes an advanced reasoning tasks, including code and math, classification and question answering, translation and multilingual proficiency, and natural language generation better than our previous state-of-the-art LLMs.\n\n[Med-PaLM](https:\u002F\u002Fsites.research.google\u002Fmed-palm\u002F) is a large language model (LLM) designed to provide high quality answers to medical questions. It harnesses the power of Google’s large language models, which we have aligned to the medical domain with a set of carefully-curated medical expert demonstrations. \n\n[Sec-PaLM](https:\u002F\u002Fcloud.google.com\u002Fblog\u002Fproducts\u002Fidentity-security\u002Frsa-google-cloud-security-ai-workbench-generative-ai) is a large language models (LLMs), that accelerate the ability to help people who are responsible for keeping their organizations safe. These new models not only give people a more natural and creative way to understand and manage security.\n\n### LLM Training Frameworks\n\n[Back to the Top](#table-of-contents)\n \n - [Alpa](https:\u002F\u002Falpa.ai\u002Findex.html) is a system for training and serving large-scale neural networks.\n - [BayLing](https:\u002F\u002Fgithub.com\u002Fictnlp\u002FBayLing) - an English\u002FChinese LLM equipped with advanced language alignment, showing superior capability in English\u002FChinese generation, instruction following and multi-turn interaction.\n - [BLOOM](https:\u002F\u002Fhuggingface.co\u002Fbigscience\u002Fbloom) - BigScience Large Open-science Open-access Multilingual Language Model [BLOOM-LoRA](https:\u002F\u002Fgithub.com\u002Flinhduongtuan\u002FBLOOM-LORA)\n - [Cerebras-GPT](https:\u002F\u002Fwww.cerebras.net\u002Fblog\u002Fcerebras-gpt-a-family-of-open-compute-efficient-large-language-models\u002F) - A Family of Open, Compute-efficient, Large Language Models.\n - [DeepSpeed](https:\u002F\u002Fwww.deepspeed.ai\u002F) is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective. \n - [FairScale](https:\u002F\u002Ffairscale.readthedocs.io\u002Fen\u002Flatest\u002Fwhat_is_fairscale.html) is a PyTorch extension library for high performance and large scale training. This library extends basic PyTorch capabilities while adding new SOTA scaling techniques.  \n - [GLM](https:\u002F\u002Fgithub.com\u002FTHUDM\u002FGLM)- GLM is a General Language Model pretrained with an autoregressive blank-filling objective and can be finetuned on various natural language understanding and generation tasks.\n - [OpenFlamingo](https:\u002F\u002Fgithub.com\u002Fmlfoundations\u002Fopen_flamingo) is an open-source framework implementation of DeepMind's [Flamingo](https:\u002F\u002Fwww.deepmind.com\u002Fblog\u002Ftackling-multiple-tasks-with-a-single-visual-language-model) for training large multimodal models.\n - [OPT](https:\u002F\u002Farxiv.org\u002Fabs\u002F2205.01068) - Open Pre-trained Transformer Language Models.\n - [StarCoder](https:\u002F\u002Fhuggingface.co\u002Fblog\u002Fstarcoder) - Hugging Face LLM for Code\n - [UltraLM](https:\u002F\u002Fgithub.com\u002Fthunlp\u002FUltraChat) - Large-scale, Informative, and Diverse Multi-round Chat Models. \n - [UL2](https:\u002F\u002Farxiv.org\u002Fabs\u002F2205.05131v1) - a unified framework for pretraining models that are universally effective across datasets and setups. \n \n \n### Tools for deploying LLM\n\n[Back to the Top](#table-of-contents)\n\n- [Agenta](https:\u002F\u002Fgithub.com\u002Fagenta-ai\u002Fagenta) -  Easily build, version, evaluate and deploy your LLM-powered apps.\n- [BentoML](https:\u002F\u002Fbentoml.com\u002F) for LLMs-based applications.\n- [CometLLM](https:\u002F\u002Fgithub.com\u002Fcomet-ml\u002Fcomet-llm) - A open-source LLMOps platform to log, manage, and visualize your LLM prompts and chains. Track prompt templates, prompt variables, prompt duration, token usage, and other metadata. Score prompt outputs and visualize chat history all within a single UI.\n- [FastChat](https:\u002F\u002Fgithub.com\u002Flm-sys\u002FFastChat) - A distributed multi-model LLM serving system with web UI and OpenAI-compatible RESTful APIs.\n- [Embedchain](https:\u002F\u002Fgithub.com\u002Fembedchain\u002Fembedchain) - Framework to create ChatGPT like bots over your dataset.\n- [IntelliServer](https:\u002F\u002Fgithub.com\u002Fintelligentnode\u002FIntelliServer) - simplifies the evaluation of LLMs by providing a unified microservice to access and test multiple AI models.\n- [Haystack](https:\u002F\u002Fhaystack.deepset.ai\u002F) - an open-source NLP framework that allows you to use LLMs and transformer-based models from Hugging Face, OpenAI and Cohere to interact with your own data. \n- [Langroid](https:\u002F\u002Fgithub.com\u002Flangroid\u002Flangroid) - Harness LLMs with Multi-Agent Programming.\n- [LangChain](https:\u002F\u002Fgithub.com\u002Fhwchase17\u002Flangchain) -  Building applications with LLMs through composability.\n- [LiteChain](https:\u002F\u002Fgithub.com\u002Frogeriochaves\u002Flitechain) - Lightweight alternative to LangChain for composing LLMs .\n- [Magentic](https:\u002F\u002Fgithub.com\u002Fjackmpcollins\u002Fmagentic) - Seamlessly integrate LLMs as Python functions.\n- [Promptfoo](https:\u002F\u002Fgithub.com\u002Ftyppo\u002Fpromptfoo) - Test your prompts. Evaluate and compare LLM outputs, catch regressions, and improve prompt quality.\n- [OpenLLM](https:\u002F\u002Fgithub.com\u002Fbentoml\u002FOpenLLM) is an open platform for operating large language models (LLMs) in production. Fine-tune, serve, deploy, and monitor any LLMs with ease.\n- [Serge](https:\u002F\u002Fgithub.com\u002Fserge-chat\u002Fserge) - a chat interface crafted with llama.cpp for running Alpaca models. No API keys, entirely self-hosted!\n- [SkyPilot](https:\u002F\u002Fgithub.com\u002Fskypilot-org\u002Fskypilot) - Run LLMs and batch jobs on any cloud. Get maximum cost savings, highest GPU availability, and managed execution -- all with a simple interface.\n- [Text Generation Inference](https:\u002F\u002Fgithub.com\u002Fhuggingface\u002Ftext-generation-inference) - A Rust, Python and gRPC server for text generation inference. Used in production at [HuggingFace](https:\u002F\u002Fhuggingface.co\u002F) to power LLMs api-inference widgets.\n- [vLLM](https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm) - A high-throughput and memory-efficient inference and serving engine for LLMs.\n\n### Running LLMs Locally\n\n[Back to the Top](#table-of-contents)\n\n * [A comprehensive guide to running Llama 2 locally](https:\u002F\u002Freplicate.com\u002Fblog\u002Frun-llama-locally)\n * [Leaderboard by lmsys.org](https:\u002F\u002Fchat.lmsys.org\u002F?leaderboard)\n * [LLM-Leaderboard](https:\u002F\u002Fgithub.com\u002FLudwigStumpp\u002Fllm-leaderboard)\n * [Open LLM Leaderboard by Hugging Face](https:\u002F\u002Fhuggingface.co\u002Fspaces\u002FHuggingFaceH4\u002Fopen_llm_leaderboard)\n * [Holistic Evaluation of Language Models (HELM)](https:\u002F\u002Fcrfm.stanford.edu\u002Fhelm\u002Flatest\u002F?groups=1)\n * [TextSynth Server Benchmarks](https:\u002F\u002Fbellard.org\u002Fts_server\u002F)\n\n[LocalAI](https:\u002F\u002Flocalai.io\u002F) is a self-hosted, community-driven, local OpenAI-compatible API. Drop-in replacement for OpenAI running LLMs on consumer-grade hardware with no GPU required. It's an API to run ggml compatible models: llama, gpt4all, rwkv, whisper, vicuna, koala, gpt4all-j, cerebras, falcon, dolly, starcoder, and many others.\n\n[llama.cpp](https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fllama.cpp) is a Port of Facebook's LLaMA model in C\u002FC++.\n\n[ollama](https:\u002F\u002Follama.ai\u002F) is a tool to get up and running with Llama 2 and other large language models locally.\n\n[LocalAI](https:\u002F\u002Flocalai.io\u002F) is a self-hosted, community-driven, local OpenAI-compatible API. Drop-in replacement for OpenAI running LLMs on consumer-grade hardware with no GPU required. It's an API to run ggml compatible models: llama, gpt4all, rwkv, whisper, vicuna, koala, gpt4all-j, cerebras, falcon, dolly, starcoder, and many others.\n \n[Serge](https:\u002F\u002Fgithub.com\u002Fserge-chat\u002Fserge) is a web interface for chatting with Alpaca through llama.cpp. Fully self-hosted & dockerized, with an easy to use API. \n\n[OpenLLM](https:\u002F\u002Fgithub.com\u002Fbentoml\u002FOpenLLM) is an open platform for operating large language models (LLMs) in production. Fine-tune, serve, deploy, and monitor any LLMs with ease.\n\n[Llama-gpt](https:\u002F\u002Fgithub.com\u002Fgetumbrel\u002Fllama-gpt) is a self-hosted, offline, ChatGPT-like chatbot. Powered by Llama 2. 100% private, with no data leaving your device. \n\n[Llama2 webui](https:\u002F\u002Fgithub.com\u002Fliltom-eth\u002Fllama2-webui) is a tool to run any Llama 2 locally with gradio UI on GPU or CPU from anywhere (Linux\u002FWindows\u002FMac). Use `llama2-wrapper` as your local llama2 backend for Generative Agents\u002FApps. \n\n[Llama2.c](https:\u002F\u002Fgithub.com\u002Fkarpathy\u002Fllama2.c) is a tool to Train the Llama 2 LLM architecture in PyTorch then inference it with one simple 700-line C file ([run.c](https:\u002F\u002Fgithub.com\u002Fkarpathy\u002Fllama2.c\u002Fblob\u002Fmaster\u002Frun.c)).\n\n[Alpaca.cpp](https:\u002F\u002Fgithub.com\u002Fantimatter15\u002Falpaca.cpp) is a fast ChatGPT-like model locally on your device. It combines the [LLaMA foundation model](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fllama) with an [open reproduction](https:\u002F\u002Fgithub.com\u002Ftloen\u002Falpaca-lora) of [Stanford Alpaca](https:\u002F\u002Fgithub.com\u002Ftatsu-lab\u002Fstanford_alpaca) a fine-tuning of the base model to obey instructions (akin to the [RLHF](https:\u002F\u002Fhuggingface.co\u002Fblog\u002Frlhf) used to train ChatGPT) and a set of modifications to [llama.cpp](https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fllama.cpp) to add a chat interface.\n\n[GPT4All](https:\u002F\u002Fgithub.com\u002Fnomic-ai\u002Fgpt4all) is an ecosystem of open-source chatbots trained on a massive collections of clean assistant data including code, stories and dialogue based on [LLaMa](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fllama).\n\n[MiniGPT-4](https:\u002F\u002Fminigpt-4.github.io\u002F) is an enhancing Vision-language Understanding with Advanced Large Language Models\n\n[LoLLMS WebUI](https:\u002F\u002Fgithub.com\u002FParisNeo\u002Flollms-webui) is a the hub for LLM (Large Language Model) models. It aims to provide a user-friendly interface to access and utilize various LLM models for a wide range of tasks. Whether you need help with writing, coding, organizing data, generating images, or seeking answers to your questions.\n\n[LM Studio](https:\u002F\u002Flmstudio.ai\u002F) is a tool to Discover, download, and run local LLMs.\n\n[Gradio Web UI](https:\u002F\u002Fgithub.com\u002Foobabooga\u002Ftext-generation-webui) is a tool for Large Language Models. Supports transformers, GPTQ, llama.cpp (ggml\u002Fgguf), Llama models. \n\n[OpenPlayground](https:\u002F\u002Fgithub.com\u002Fnat\u002Fopenplayground) is a playfround for running ChatGPT-like models locally on your device.\n\n[Vicuna](https:\u002F\u002Fvicuna.lmsys.org\u002F) is an open source chatbot trained by fine tuning LLaMA. It apparently achieves more than 90% quality of chatgpt and costs $300 to train.\n\n[Yeagar ai](https:\u002F\u002Fgithub.com\u002Fyeagerai\u002Fyeagerai-agent) is a Langchain Agent creator designed to help you build, prototype, and deploy AI-powered agents with ease.\n\n[KoboldCpp](https:\u002F\u002Fgithub.com\u002FLostRuins\u002Fkoboldcpp) is an easy-to-use AI text-generation software for GGML models. It's a single self contained distributable from Concedo, that builds off llama.cpp, and adds a versatile Kobold API endpoint, additional format support, backward compatibility, as well as a fancy UI with persistent stories, editing tools, save formats, memory, world info, author's note, characters, and scenarios.\n\n# Algorithms\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n[Fuzzy logic](https:\u002F\u002Fwww.investopedia.com\u002Fterms\u002Ff\u002Ffuzzy-logic.asp) is a heuristic approach that allows for more advanced decision-tree processing and better integration with rules-based programming.\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_751e3caad00a.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**Architecture of a Fuzzy Logic System. Source: [ResearchGate](https:\u002F\u002Fwww.researchgate.net\u002Ffigure\u002FArchitecture-of-a-fuzzy-logic-system_fig2_309452475)**\n\n[Support Vector Machine (SVM)](https:\u002F\u002Fweb.stanford.edu\u002F~hastie\u002FMOOC-Slides\u002Fsvm.pdf) is a supervised machine learning model that uses classification algorithms for two-group classification problems.\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_db9febb4b018.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**Support Vector Machine (SVM). Source:[OpenClipArt](https:\u002F\u002Fopenclipart.org\u002Fdetail\u002F182977\u002Fsvm-support-vector-machines)**\n\n[Neural networks](https:\u002F\u002Fwww.ibm.com\u002Fcloud\u002Flearn\u002Fneural-networks) are a subset of machine learning and are at the heart of deep learning algorithms. The name\u002Fstructure is inspired by the human brain copying the process that biological neurons\u002Fnodes signal to one another.\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_4927bf734ef7.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**Deep neural network. Source: [IBM](https:\u002F\u002Fwww.ibm.com\u002Fcloud\u002Flearn\u002Fneural-networks)**\n\n[Convolutional Neural Networks (R-CNN)](https:\u002F\u002Fstanford.edu\u002F~shervine\u002Fteaching\u002Fcs-230\u002Fcheatsheet-convolutional-neural-networks) is an object detection algorithm that first segments the image to find potential relevant bounding boxes and then run the detection algorithm to find most probable objects in those bounding boxes.\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_8f5b375d6e6c.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**Convolutional Neural Networks. Source:[CS231n](https:\u002F\u002Fcs231n.github.io\u002Fconvolutional-networks\u002F#conv)**\n\n[Recurrent neural networks (RNNs)](https:\u002F\u002Fwww.ibm.com\u002Fcloud\u002Flearn\u002Frecurrent-neural-networks) is a type of artificial neural network which uses sequential data or time series data.\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_ddd80740aee5.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**Recurrent Neural Networks. Source: [Slideteam](https:\u002F\u002Fwww.slideteam.net\u002Frecurrent-neural-networks-rnns-ppt-powerpoint-presentation-file-templates.html)**\n\n[Multilayer Perceptrons (MLPs)](https:\u002F\u002Fdeepai.org\u002Fmachine-learning-glossary-and-terms\u002Fmultilayer-perceptron) is multi-layer neural networks composed of multiple layers of [perceptrons](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FPerceptron) with a threshold activation.\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_837d266dc173.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**Multilayer Perceptrons. Source: [DeepAI](https:\u002F\u002Fdeepai.org\u002Fmachine-learning-glossary-and-terms\u002Fmultilayer-perceptron)**\n\n[Random forest](https:\u002F\u002Fwww.ibm.com\u002Fcloud\u002Flearn\u002Frandom-forest) is a commonly-used machine learning algorithm, which combines the output of multiple decision trees to reach a single result. A decision tree in a forest cannot be pruned for sampling and therefore, prediction selection. Its ease of use and flexibility have fueled its adoption, as it handles both classification and regression problems.\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_5c55812f2c05.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**Random forest. Source: [wikimedia](https:\u002F\u002Fcommunity.tibco.com\u002Fwiki\u002Frandom-forest-template-tibco-spotfirer-wiki-page)**\n\n[Decision trees](https:\u002F\u002Fwww.cs.cmu.edu\u002F~bhiksha\u002Fcourses\u002F10-601\u002Fdecisiontrees\u002F) are tree-structured models for classification and regression.\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_ee5160d69cb4.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n***Decision Trees. Source: [CMU](http:\u002F\u002Fwww.cs.cmu.edu\u002F~bhiksha\u002Fcourses\u002F10-601\u002Fdecisiontrees\u002F)*\n\n[Naive Bayes](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FNaive_Bayes_classifier) is a machine learning algorithm that is used solved calssification problems. It's based on applying [Bayes' theorem](https:\u002F\u002Fwww.mathsisfun.com\u002Fdata\u002Fbayes-theorem.html) with strong independence assumptions between the features.\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_50971225cc23.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**Bayes' theorem. Source:[mathisfun](https:\u002F\u002Fwww.mathsisfun.com\u002Fdata\u002Fbayes-theorem.html)**\n\n# PyTorch Development\n\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_b48222faa05e.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## PyTorch Learning Resources\n\n[PyTorch](https:\u002F\u002Fpytorch.org) is an open-source deep learning framework that accelerates the path from research to production, used for applications such as computer vision and natural language processing. PyTorch is developed by [Facebook's AI Research](https:\u002F\u002Fai.facebook.com\u002Fresearch\u002F) lab.\n\n[Getting Started with PyTorch](https:\u002F\u002Fpytorch.org\u002Fget-started\u002Flocally\u002F)\n\n[PyTorch Documentation](https:\u002F\u002Fpytorch.org\u002Fdocs\u002Fstable\u002Findex.html)\n\n[PyTorch Discussion Forum](https:\u002F\u002Fdiscuss.pytorch.org\u002F)\n\n[Top Pytorch Courses Online | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=pytorch&page=1)\n\n[Top Pytorch Courses Online | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002FPyTorch\u002F)\n\n[Learn PyTorch with Online Courses and Classes | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fpytorch)\n\n[PyTorch Fundamentals - Learn | Microsoft Docs](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Flearn\u002Fpaths\u002Fpytorch-fundamentals\u002F)\n\n[Intro to Deep Learning with PyTorch | Udacity ](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fdeep-learning-pytorch--ud188)\n\n[PyTorch Development in Visual Studio Code](https:\u002F\u002Fcode.visualstudio.com\u002Fdocs\u002Fdatascience\u002Fpytorch-support)\n\n[PyTorch on Azure - Deep Learning with PyTorch | Microsoft Azure](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fdevelop\u002Fpytorch\u002F)\n\n[PyTorch - Azure Databricks | Microsoft Docs](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fdatabricks\u002Fapplications\u002Fmachine-learning\u002Ftrain-model\u002Fpytorch)\n\n[Deep Learning with PyTorch | Amazon Web Services (AWS)](https:\u002F\u002Faws.amazon.com\u002Fpytorch\u002F)\n\n[Getting started with PyTorch on Google Cloud](https:\u002F\u002Fcloud.google.com\u002Fai-platform\u002Ftraining\u002Fdocs\u002Fgetting-started-pytorch)\n\n## PyTorch Tools, Libraries, and Frameworks\n\n[PyTorch Mobile](https:\u002F\u002Fpytorch.org\u002Fmobile\u002Fhome\u002F) is an end-to-end ML workflow from Training to Deployment for iOS and Android mobile devices.\n\n[TorchScript](https:\u002F\u002Fpytorch.org\u002Fdocs\u002Fstable\u002Fjit.html) is a way to create serializable and optimizable models from PyTorch code. This allows any TorchScript program to be saved from a Python process and loaded in a process where there is no Python dependency.\n\n[TorchServe](https:\u002F\u002Fpytorch.org\u002Fserve\u002F) is a flexible and easy to use tool for serving PyTorch models.\n\n[Keras](https:\u002F\u002Fkeras.io) is a high-level neural networks API, written in Python and capable of running on top of TensorFlow, CNTK, or Theano.It was developed with a focus on enabling fast experimentation. It is capable of running on top of TensorFlow, Microsoft Cognitive Toolkit, R, Theano, or PlaidML.\n\n[ONNX Runtime](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fonnxruntime) is a cross-platform, high performance ML inferencing and training accelerator. It supports models from deep learning frameworks such as PyTorch and TensorFlow\u002FKeras as well as classical machine learning libraries such as scikit-learn, LightGBM, XGBoost, etc.\n\n[Kornia](https:\u002F\u002Fkornia.github.io\u002F) is a differentiable computer vision library that consists of a set of routines and differentiable modules to solve generic CV (Computer Vision) problems.\n\n[PyTorch-NLP](https:\u002F\u002Fpytorchnlp.readthedocs.io\u002Fen\u002Flatest\u002F) is a library for Natural Language Processing (NLP) in Python. It’s built with the very latest research in mind, and was designed from day one to support rapid prototyping. PyTorch-NLP comes with pre-trained embeddings, samplers, dataset loaders, metrics, neural network modules and text encoders.\n\n[Ignite](https:\u002F\u002Fpytorch.org\u002Fignite) is a high-level library to help with training and evaluating neural networks in PyTorch flexibly and transparently.\n\n[Hummingbird](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fhummingbird) is a library for compiling trained traditional ML models into tensor computations. It allows users to seamlessly leverage neural network frameworks (such as PyTorch) to accelerate traditional ML models.\n\n[Deep Graph Library (DGL)](https:\u002F\u002Fwww.dgl.ai\u002F) is a Python package built for easy implementation of graph neural network model family, on top of PyTorch and other frameworks.\n\n[TensorLy](http:\u002F\u002Ftensorly.org\u002Fstable\u002Fhome.html) is a high level API for tensor methods and deep tensorized neural networks in Python that aims to make tensor learning simple.\n\n[GPyTorch](https:\u002F\u002Fcornellius-gp.github.io\u002F) is a Gaussian process library implemented using PyTorch, designed for creating scalable, flexible Gaussian process models.\n\n[Poutyne](https:\u002F\u002Fpoutyne.org\u002F) is a Keras-like framework for PyTorch and handles much of the boilerplating code needed to train neural networks.\n\n[Forte](https:\u002F\u002Fgithub.com\u002Fasyml\u002Fforte\u002Ftree\u002Fmaster\u002Fdocs) is a toolkit for building NLP pipelines featuring composable components, convenient data interfaces, and cross-task interaction.\n\n[TorchMetrics](https:\u002F\u002Fgithub.com\u002FPyTorchLightning\u002Fmetrics) is a Machine learning metrics for distributed, scalable PyTorch applications.\n\n[Captum](https:\u002F\u002Fcaptum.ai\u002F) is an open source, extensible library for model interpretability built on PyTorch.\n\n[Transformer](https:\u002F\u002Fgithub.com\u002Fhuggingface\u002Ftransformers) is a State-of-the-art Natural Language Processing for Pytorch, TensorFlow, and JAX.\n\n[Hydra](https:\u002F\u002Fhydra.cc) is a framework for elegantly configuring complex applications.\n\n[Accelerate](https:\u002F\u002Fhuggingface.co\u002Fdocs\u002Faccelerate) is a simple way to train and use PyTorch models with multi-GPU, TPU, mixed-precision.\n\n[Ray](https:\u002F\u002Fgithub.com\u002Fray-project\u002Fray) is a fast and simple framework for building and running distributed applications.\n\n[ParlAI](http:\u002F\u002Fparl.ai\u002F) is a unified platform for sharing, training, and evaluating dialog models across many tasks.\n\n[PyTorchVideo](https:\u002F\u002Fpytorchvideo.org\u002F) is a deep learning library for video understanding research. Hosts various video-focused models, datasets, training pipelines and more.\n\n[Opacus](https:\u002F\u002Fopacus.ai\u002F) is a library that enables training PyTorch models with Differential Privacy.\n\n[PyTorch Lightning](https:\u002F\u002Fgithub.com\u002FwilliamFalcon\u002Fpytorch-lightning) is a Keras-like ML library for PyTorch. It leaves core training and validation logic to you and automates the rest.\n\n[PyTorch Geometric Temporal](https:\u002F\u002Fgithub.com\u002Fbenedekrozemberczki\u002Fpytorch_geometric_temporal) is a temporal (dynamic) extension library for PyTorch Geometric.\n\n[PyTorch Geometric](https:\u002F\u002Fgithub.com\u002Frusty1s\u002Fpytorch_geometric) is a library for deep learning on irregular input data such as graphs, point clouds, and manifolds.\n\n[Raster Vision](https:\u002F\u002Fdocs.rastervision.io\u002F) is an open source framework for deep learning on satellite and aerial imagery.\n\n[CrypTen](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002FCrypTen) is a framework for Privacy Preserving ML. Its goal is to make secure computing techniques accessible to ML practitioners.\n\n[Optuna](https:\u002F\u002Foptuna.org\u002F) is an open source hyperparameter optimization framework to automate hyperparameter search.\n\n[Pyro](http:\u002F\u002Fpyro.ai\u002F) is a universal probabilistic programming language (PPL) written in Python and supported by PyTorch on the backend.\n\n[Albumentations](https:\u002F\u002Fgithub.com\u002Falbu\u002Falbumentations) is a fast and extensible image augmentation library for different CV tasks like classification, segmentation, object detection and pose estimation.\n\n[Skorch](https:\u002F\u002Fgithub.com\u002Fskorch-dev\u002Fskorch) is a high-level library for PyTorch that provides full scikit-learn compatibility.\n\n[MMF](https:\u002F\u002Fmmf.sh\u002F) is a modular framework for vision & language multimodal research from Facebook AI Research (FAIR).\n\n[AdaptDL](https:\u002F\u002Fgithub.com\u002Fpetuum\u002Fadaptdl) is a resource-adaptive deep learning training and scheduling framework.\n\n[Polyaxon](https:\u002F\u002Fgithub.com\u002Fpolyaxon\u002Fpolyaxon) is a platform for building, training, and monitoring large-scale deep learning applications.\n\n[TextBrewer](http:\u002F\u002Ftextbrewer.hfl-rc.com\u002F) is a PyTorch-based knowledge distillation toolkit for natural language processing\n\n[AdverTorch](https:\u002F\u002Fgithub.com\u002FBorealisAI\u002Fadvertorch) is a toolbox for adversarial robustness research. It contains modules for generating adversarial examples and defending against attacks.\n\n[NeMo](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNeMo) is a a toolkit for conversational AI.\n\n[ClinicaDL](https:\u002F\u002Fclinicadl.readthedocs.io\u002F) is a framework for reproducible classification of Alzheimer's Disease\n\n[Stable Baselines3 (SB3)](https:\u002F\u002Fgithub.com\u002FDLR-RM\u002Fstable-baselines3) is a set of reliable implementations of reinforcement learning algorithms in PyTorch.\n\n[TorchIO](https:\u002F\u002Fgithub.com\u002Ffepegar\u002Ftorchio) is a set of tools to efficiently read, preprocess, sample, augment, and write 3D medical images in deep learning applications written in PyTorch.\n\n[PySyft](https:\u002F\u002Fgithub.com\u002FOpenMined\u002FPySyft) is a Python library for encrypted, privacy preserving deep learning.\n\n[Flair](https:\u002F\u002Fgithub.com\u002FflairNLP\u002Fflair) is a very simple framework for state-of-the-art natural language processing (NLP).\n\n[Glow](https:\u002F\u002Fgithub.com\u002Fpytorch\u002Fglow) is a ML compiler that accelerates the performance of deep learning frameworks on different hardware platforms.\n\n[FairScale](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Ffairscale) is a PyTorch extension library for high performance and large scale training on one or multiple machines\u002Fnodes.\n\n[MONAI](https:\u002F\u002Fmonai.io\u002F) is a deep learning framework that provides domain-optimized foundational capabilities for developing healthcare imaging training workflows.\n\n[PFRL](https:\u002F\u002Fgithub.com\u002Fpfnet\u002Fpfrl) is a deep reinforcement learning library that implements various state-of-the-art deep reinforcement algorithms in Python using PyTorch.\n\n[Einops](https:\u002F\u002Fgithub.com\u002Farogozhnikov\u002Feinops) is a flexible and powerful tensor operations for readable and reliable code.\n\n[PyTorch3D](https:\u002F\u002Fpytorch3d.org\u002F) is a deep learning library that provides efficient, reusable components for 3D Computer Vision research with PyTorch.\n\n[Ensemble Pytorch](https:\u002F\u002Fensemble-pytorch.readthedocs.io\u002F) is a unified ensemble framework for PyTorch to improve the performance and robustness of your deep learning model.\n\n[Lightly](https:\u002F\u002Fgithub.com\u002Flightly-ai\u002Flightly) is a computer vision framework for self-supervised learning.\n\n[Higher](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fhigher) is a library which facilitates the implementation of arbitrarily complex gradient-based meta-learning algorithms and nested optimisation loops with near-vanilla PyTorch.\n\n[Horovod](http:\u002F\u002Fhorovod.ai\u002F) is a distributed training library for deep learning frameworks. Horovod aims to make distributed DL fast and easy to use.\n\n[PennyLane](https:\u002F\u002Fpennylane.ai\u002F) is a library for quantum ML, automatic differentiation, and optimization of hybrid quantum-classical computations.\n\n[Detectron2](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fdetectron2) is FAIR's next-generation platform for object detection and segmentation.\n\n[Fastai](https:\u002F\u002Fdocs.fast.ai\u002F) is a library that simplifies training fast and accurate neural nets using modern best practices.\n\n# TensorFlow Development\n\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_809fa9630256.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## TensorFlow Learning Resources\n\n[TensorFlow](https:\u002F\u002Fwww.tensorflow.org) is an end-to-end open source platform for machine learning. It has a comprehensive, flexible ecosystem of tools, libraries and community resources that lets researchers push the state-of-the-art in ML and developers easily build and deploy ML powered applications.\n\n[Getting Started with TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002Flearn)\n\n[TensorFlow Tutorials](https:\u002F\u002Fwww.tensorflow.org\u002Ftutorials\u002F)\n\n[TensorFlow Developer Certificate | TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002Fcertificate)\n\n[TensorFlow Community](https:\u002F\u002Fwww.tensorflow.org\u002Fcommunity\u002F)\n\n[TensorFlow Models & Datasets](https:\u002F\u002Fwww.tensorflow.org\u002Fresources\u002Fmodels-datasets)\n\n[TensorFlow Cloud](https:\u002F\u002Fwww.tensorflow.org\u002Fcloud)\n\n[Machine learning education | TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002Fresources\u002Flearn-ml)\n\n[Top Tensorflow Courses Online | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=tensorflow)\n\n[Top Tensorflow Courses Online | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourses\u002Fsearch\u002F?src=ukw&q=tensorflow)\n\n[Deep Learning with TensorFlow | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fdeep-learning-with-tensorflow-certification-training\u002F)\n\n[Deep Learning with Tensorflow | edX](https:\u002F\u002Fwww.edx.org\u002Fcourse\u002Fdeep-learning-with-tensorflow)\n\n[Intro to TensorFlow for Deep Learning | Udacity ](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fintro-to-tensorflow-for-deep-learning--ud187)\n\n[Intro to TensorFlow: Machine Learning Crash Course | Google Developers](https:\u002F\u002Fdevelopers.google.com\u002Fmachine-learning\u002Fcrash-course\u002Ffirst-steps-with-tensorflow\u002Ftoolkit)\n\n[Train and deploy a TensorFlow model - Azure Machine Learning](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fmachine-learning\u002Fhow-to-train-tensorflow)\n\n[Apply machine learning models in Azure Functions with Python and TensorFlow | Microsoft Azure](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fazure-functions\u002Ffunctions-machine-learning-tensorflow?tabs=bash)\n\n[Deep Learning with TensorFlow | Amazon Web Services (AWS)](https:\u002F\u002Faws.amazon.com\u002Ftensorflow\u002F)\n\n[TensorFlow - Amazon EMR | AWS Documentation](https:\u002F\u002Fdocs.aws.amazon.com\u002Femr\u002Flatest\u002FReleaseGuide\u002Femr-tensorflow.html)\n\n[TensorFlow Enterprise | Google Cloud](https:\u002F\u002Fcloud.google.com\u002Ftensorflow-enterprise\u002F)\n\n## TensorFlow Tools, Libraries, and Frameworks\n\n[TensorFlow Lite](https:\u002F\u002Fwww.tensorflow.org\u002Flite\u002F) is an open source deep learning framework for deploying machine learning models on mobile and IoT devices.\n\n[TensorFlow.js](https:\u002F\u002Fwww.tensorflow.org\u002Fjs) is a JavaScript Library that lets you develop or execute ML models in JavaScript, and use ML directly in the browser client side, server side via Node.js, mobile native via React Native, desktop native via Electron, and even on IoT devices via Node.js on Raspberry Pi.\n\n[Tensorflow_macOS](https:\u002F\u002Fgithub.com\u002Fapple\u002Ftensorflow_macos) is a Mac-optimized version of TensorFlow and TensorFlow Addons for macOS 11.0+ accelerated using Apple's ML Compute framework.\n\n[Google Colaboratory](https:\u002F\u002Fcolab.sandbox.google.com\u002Fnotebooks\u002Fwelcome.ipynb) is a free Jupyter notebook environment that requires no setup and runs entirely in the cloud, allowing you to execute TensorFlow code in your browser with a single click.\n\n[What-If Tool](https:\u002F\u002Fpair-code.github.io\u002Fwhat-if-tool\u002F) is a tool for code-free probing of machine learning models, useful for model understanding, debugging, and fairness. Available in TensorBoard and jupyter or colab notebooks.\n\n[TensorBoard](https:\u002F\u002Fwww.tensorflow.org\u002Ftensorboard) is a suite of visualization tools to understand, debug, and optimize TensorFlow programs.\n\n[Keras](https:\u002F\u002Fkeras.io) is a high-level neural networks API, written in Python and capable of running on top of TensorFlow, CNTK, or Theano.It was developed with a focus on enabling fast experimentation. It is capable of running on top of TensorFlow, Microsoft Cognitive Toolkit, R, Theano, or PlaidML.\n\n[XLA (Accelerated Linear Algebra)](https:\u002F\u002Fwww.tensorflow.org\u002Fxla) is a domain-specific compiler for linear algebra that optimizes TensorFlow computations. The results are improvements in speed, memory usage, and portability on server and mobile platforms.\n\n[ML Perf](https:\u002F\u002Fmlperf.org\u002F) is a broad ML benchmark suite for measuring performance of ML software frameworks, ML hardware accelerators, and ML cloud platforms.\n\n[TensorFlow Playground](https:\u002F\u002Fplayground.tensorflow.org\u002F#activation=tanh&batchSize=10&dataset=circle&regDataset=reg-plane&learningRate=0.03&regularizationRate=0&noise=0&networkShape=4,2&seed=0.04620&showTestData=false&discretize=false&percTrainData=50&x=true&y=true&xTimesY=false&xSquared=false&ySquared=false&cosX=false&sinX=false&cosY=false&sinY=false&collectStats=false&problem=classification&initZero=false&hideText=false) is an development environment to tinker around with a neural network in your browser.\n\n[TPU Research Cloud (TRC)](https:\u002F\u002Fsites.research.google\u002Ftrc\u002F) is a program enables researchers to apply for access to a cluster of more than 1,000 Cloud TPUs at no charge to help them accelerate the next wave of research breakthroughs.\n\n[MLIR](https:\u002F\u002Fwww.tensorflow.org\u002Fmlir) is a new intermediate representation and compiler framework.\n\n[Lattice](https:\u002F\u002Fwww.tensorflow.org\u002Flattice) is a library for flexible, controlled and interpretable ML solutions with common-sense shape constraints.\n\n[TensorFlow Hub](https:\u002F\u002Fwww.tensorflow.org\u002Fhub) is a library for reusable machine learning. Download and reuse the latest trained models with a minimal amount of code.\n\n[TensorFlow Cloud](https:\u002F\u002Fwww.tensorflow.org\u002Fcloud) is a library to connect your local environment to Google Cloud.\n\n[TensorFlow Model Optimization Toolkit](https:\u002F\u002Fwww.tensorflow.org\u002Fmodel_optimization) is a suite of tools for optimizing ML models for deployment and execution.\n\n[TensorFlow Recommenders](https:\u002F\u002Fwww.tensorflow.org\u002Frecommenders) is a library for building recommender system models.\n\n[TensorFlow Text](https:\u002F\u002Fwww.tensorflow.org\u002Ftext) is a collection of text- and NLP-related classes and ops ready to use with TensorFlow 2.\n\n[TensorFlow Graphics](https:\u002F\u002Fwww.tensorflow.org\u002Fgraphics) is a library of computer graphics functionalities ranging from cameras, lights, and materials to renderers.\n\n[TensorFlow Federated](https:\u002F\u002Fwww.tensorflow.org\u002Ffederated) is an open source framework for machine learning and other computations on decentralized data.\n\n[TensorFlow Probability](https:\u002F\u002Fwww.tensorflow.org\u002Fprobability) is a library for probabilistic reasoning and statistical analysis.\n\n[Tensor2Tensor](https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Ftensor2tensor) is a library of deep learning models and datasets designed to make deep learning more accessible and accelerate ML research.\n\n[TensorFlow Privacy](https:\u002F\u002Fwww.tensorflow.org\u002Fresponsible_ai\u002Fprivacy) is a Python library that includes implementations of TensorFlow optimizers for training machine learning models with differential privacy.\n\n[TensorFlow Ranking](https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Franking) is a library for Learning-to-Rank (LTR) techniques on the TensorFlow platform.\n\n[TensorFlow Agents](https:\u002F\u002Fwww.tensorflow.org\u002Fagents) is a library for reinforcement learning in TensorFlow.\n\n[TensorFlow Addons](https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Faddons) is a repository of contributions that conform to well-established API patterns, but implement new functionality not available in core TensorFlow, maintained by [SIG Addons](https:\u002F\u002Fgroups.google.com\u002Fa\u002Ftensorflow.org\u002Fg\u002Faddons). TensorFlow natively supports a large number of operators, layers, metrics, losses, and optimizers.\n\n[TensorFlow I\u002FO](https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Fio) is a Dataset, streaming, and file system extensions, maintained by SIG IO.\n\n[TensorFlow Quantum](https:\u002F\u002Fwww.tensorflow.org\u002Fquantum) is a quantum machine learning library for rapid prototyping of hybrid quantum-classical ML models.\n\n[Dopamine](https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fdopamine) is a research framework for fast prototyping of reinforcement learning algorithms.\n\n[TRFL](https:\u002F\u002Fdeepmind.com\u002Fblog\u002Ftrfl\u002F) is a library for reinforcement learning building blocks created by DeepMind.\n\n[Mesh TensorFlow](https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Fmesh) is a language for distributed deep learning, capable of specifying a broad class of distributed tensor computations.\n\n[RaggedTensors](https:\u002F\u002Fwww.tensorflow.org\u002Fguide\u002Fragged_tensor) is an API that makes it easy to store and manipulate data with non-uniform shape, including text (words, sentences, characters), and batches of variable length.\n\n[Unicode Ops](https:\u002F\u002Fwww.tensorflow.org\u002Ftutorials\u002Fload_data\u002Funicode) is an API that Supports working with Unicode text directly in TensorFlow.\n\n[Magenta](https:\u002F\u002Fmagenta.tensorflow.org\u002F) is a research project exploring the role of machine learning in the process of creating art and music.\n\n[Nucleus](https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fnucleus) is a library of Python and C++ code designed to make it easy to read, write and analyze data in common genomics file formats like SAM and VCF.\n\n[Sonnet](https:\u002F\u002Fgithub.com\u002Fdeepmind\u002Fsonnet) is a library from DeepMind for constructing neural networks.\n\n[Neural Structured Learning](https:\u002F\u002Fwww.tensorflow.org\u002Fneural_structured_learning) is a learning framework to train neural networks by leveraging structured signals in addition to feature inputs.\n\n[Model Remediation](https:\u002F\u002Fwww.tensorflow.org\u002Fresponsible_ai\u002Fmodel_remediation) is a library to help create and train models in a way that reduces or eliminates user harm resulting from underlying performance biases.\n\n[Fairness Indicators](https:\u002F\u002Fwww.tensorflow.org\u002Fresponsible_ai\u002Ffairness_indicators\u002Fguide) is a library that enables easy computation of commonly-identified fairness metrics for binary and multiclass classifiers.\n\n[Decision Forests](https:\u002F\u002Fwww.tensorflow.org\u002Fdecision_forests) is a State-of-the-art algorithms for training, serving and interpreting models that use decision forests for classification, regression and ranking.\n\n# Core ML Development\n\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_602cc9e929a7.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## Core ML Learning Resources\n\n[Core ML](https:\u002F\u002Fdeveloper.apple.com\u002Fdocumentation\u002Fcoreml) is an Apple framework for integrating machine learning models into apps running on Apple devices (including iOS, watchOS, macOS, and tvOS). Core ML introduces a public file format (.mlmodel) for a broad set of ML methods including deep neural networks (both convolutional and recurrent), tree ensembles with boosting, and generalized linear models. Models in this format can be directly integrated into apps through Xcode.\n\n[Introduction to Core ML](https:\u002F\u002Fcoremltools.readme.io\u002Fdocs)\n\n[Integrating a Core ML Model into your App](https:\u002F\u002Fdeveloper.apple.com\u002Fdocumentation\u002Fcoreml\u002Fintegrating_a_core_ml_model_into_your_app)\n\n[Core ML Models](https:\u002F\u002Fdeveloper.apple.com\u002Fmachine-learning\u002Fmodels\u002F)\n\n[Core ML API Reference](https:\u002F\u002Fapple.github.io\u002Fcoremltools\u002Findex.html)\n\n[Core ML Specification](https:\u002F\u002Fapple.github.io\u002Fcoremltools\u002Fmlmodel\u002Findex.html)\n\n[Apple Developer Forums for Core ML](https:\u002F\u002Fdeveloper.apple.com\u002Fforums\u002Ftags\u002Fcore-ml)\n\n[Top Core ML Courses Online | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002FCore-ML\u002F)\n\n[Top Core ML Courses Online | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=core%20ml)\n\n[IBM Watson Services for Core ML | IBM](https:\u002F\u002Fwww.ibm.com\u002Fwatson\u002Fstories\u002Fcoreml)\n\n[Generate Core ML assets using IBM Maximo Visual Inspection | IBM](https:\u002F\u002Fdeveloper.ibm.com\u002Ftechnologies\u002Fiot\u002Ftutorials\u002Fibm-maximo-visual-inspection-apple-devices\u002F)\n\n## Core ML Tools, Libraries, and Frameworks\n\n[Core ML tools](https:\u002F\u002Fcoremltools.readme.io\u002F) is a project that contains supporting tools for Core ML model conversion, editing, and validation.\n\n[Create ML](https:\u002F\u002Fdeveloper.apple.com\u002Fmachine-learning\u002Fcreate-ml\u002F) is a tool that provides new ways of training machine learning models on your Mac. It takes the complexity out of model training while producing powerful Core ML models.\n\n[Tensorflow_macOS](https:\u002F\u002Fgithub.com\u002Fapple\u002Ftensorflow_macos) is a Mac-optimized version of TensorFlow and TensorFlow Addons for macOS 11.0+ accelerated using Apple's ML Compute framework.\n\n[Apple Vision](https:\u002F\u002Fdeveloper.apple.com\u002Fdocumentation\u002Fvision) is a framework that performs face and face landmark detection, text detection, barcode recognition, image registration, and general feature tracking. Vision also allows the use of custom Core ML models for tasks like classification or object detection.\n\n[Keras](https:\u002F\u002Fkeras.io) is a high-level neural networks API, written in Python and capable of running on top of TensorFlow, CNTK, or Theano.It was developed with a focus on enabling fast experimentation. It is capable of running on top of TensorFlow, Microsoft Cognitive Toolkit, R, Theano, or PlaidML.\n\n[XGBoost](https:\u002F\u002Fxgboost.readthedocs.io\u002F) is an optimized distributed gradient boosting library designed to be highly efficient, flexible and portable. It implements machine learning algorithms under the Gradient Boosting framework. XGBoost provides a parallel tree boosting (also known as GBDT, GBM) that solve many data science problems in a fast and accurate way. It supports distributed training on multiple machines, including AWS, GCE, Azure, and Yarn clusters. Also, it can be integrated with Flink, Spark and other cloud dataflow systems.\n\n[LIBSVM](https:\u002F\u002Fwww.csie.ntu.edu.tw\u002F~cjlin\u002Flibsvm\u002F) is an integrated software for support vector classification, (C-SVC, nu-SVC), regression (epsilon-SVR, nu-SVR) and distribution estimation (one-class SVM). It supports multi-class classification.\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) is a simple and efficient tool for data mining and data analysis. It is built on NumPy,SciPy, and mathplotlib.\n\n[Xcode](https:\u002F\u002Fdeveloper.apple.com\u002Fxcode\u002F) includes everything developers need to create great applications for Mac, iPhone, iPad, Apple TV, and Apple Watch. Xcode provides developers a unified workflow for user interface design, coding, testing, and debugging. Xcode is built as an Universal app that runs 100% natively on Intel-based CPUs and Apple Silicon. It includes a unified macOS SDK that features all the frameworks, compilers, debuggers, and other tools you need to build apps that run natively on Apple Silicon and the Intel x86_64 CPU.\n\n[SwiftUI](https:\u002F\u002Fdeveloper.apple.com\u002Fdocumentation\u002Fswiftui) is a user interface toolkit that provides views, controls, and layout structures for declaring your app's user interface. The SwiftUI framework provides event handlers for delivering taps, gestures, and other types of input to your application.\n\n[UIKit](https:\u002F\u002Fdeveloper.apple.com\u002Fdocumentation\u002Fuikit) is a framework provides the required infrastructure for your iOS or tvOS apps. It provides the window and view architecture for implementing your interface, the event handling infrastructure for delivering Multi-Touch and other types of input to your app, and the main run loop needed to manage interactions among the user, the system, and your app.\n\n[AppKit](https:\u002F\u002Fdeveloper.apple.com\u002Fdocumentation\u002Fappkit) is a graphical user interface toolkit that contains all the objects you need to implement the user interface for a macOS app such as windows, panels, buttons, menus, scrollers, and text fields, and it handles all the details for you as it efficiently draws on the screen, communicates with hardware devices and screen buffers, clears areas of the screen before drawing, and clips views.\n\n[ARKit](https:\u002F\u002Fdeveloper.apple.com\u002Faugmented-reality\u002Farkit\u002F) is a set set of software development tools to enable developers to build augmented-reality apps for iOS developed by Apple. The latest version ARKit 3.5 takes advantage of the new LiDAR Scanner and depth sensing system on iPad Pro(2020) to support a new generation of AR apps that use Scene Geometry for enhanced scene understanding and object occlusion.\n\n[RealityKit](https:\u002F\u002Fdeveloper.apple.com\u002Fdocumentation\u002Frealitykit) is a framework to implement high-performance 3D simulation and rendering with information provided by the ARKit framework to seamlessly integrate virtual objects into the real world.\n\n[SceneKit](https:\u002F\u002Fdeveloper.apple.com\u002Fscenekit\u002F) is a high-level 3D graphics framework that helps you create 3D animated scenes and effects in your iOS apps.\n\n[Instruments](https:\u002F\u002Fhelp.apple.com\u002Finstruments\u002Fmac\u002Fcurrent\u002F#\u002Fdev7b09c84f5) is a powerful and flexible performance-analysis and testing tool that’s part of the Xcode tool set. It’s designed to help you profile your iOS, watchOS, tvOS, and macOS apps, processes, and devices in order to better understand and optimize their behavior and performance.\n\n[Cocoapods](https:\u002F\u002Fcocoapods.org\u002F) is a dependency manager for Swift and Objective-C used in Xcode projects by specifying the dependencies for your project in a simple text file. CocoaPods then recursively resolves dependencies between libraries, fetches source code for all dependencies, and creates and maintains an Xcode workspace to build your project.\n\n[AppCode](https:\u002F\u002Fwww.jetbrains.com\u002Fobjc\u002F) is constantly monitoring the quality of your code. It warns you of errors and smells and suggests quick-fixes to resolve them automatically. AppCode provides lots of code inspections for Objective-C, Swift, C\u002FC++, and a number of code inspections for other supported languages.\n\n\n# Deep Learning Development\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_d9156e28e925.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## Deep Learning Learning Resources\n\n[Deep Learning](https:\u002F\u002Fwww.ibm.com\u002Fcloud\u002Flearn\u002Fdeep-learning) is a subset of machine learning, which is essentially a neural network with three or more layers. These neural networks attempt to simulate the behavior of the human brain,though, far from matching its ability. This allows the neural networks to \"learn\" from large amounts of data. The Learning can be [supervised](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FSupervised_learning), [semi-supervised](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FSemi-supervised_learning) or [unsupervised](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FUnsupervised_learning).\n\n[Deep Learning Online Courses | NVIDIA](https:\u002F\u002Fwww.nvidia.com\u002Fen-us\u002Ftraining\u002Fonline\u002F)\n\n[Top Deep Learning Courses Online | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=deep%20learning)\n\n[Top Deep Learning Courses Online | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fdeep-learning\u002F)\n\n[Learn Deep Learning with Online Courses and Lessons | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fdeep-learning)\n\n[Deep Learning Online Course Nanodegree | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fdeep-learning-nanodegree--nd101)\n\n[Machine Learning Course by Andrew Ng | Coursera](https:\u002F\u002Fwww.coursera.org\u002Flearn\u002Fmachine-learning?)\n\n[Machine Learning Engineering for Production (MLOps) course by Andrew Ng | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fspecializations\u002Fmachine-learning-engineering-for-production-mlops)\n\n[Data Science: Deep Learning and Neural Networks in Python | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fdata-science-deep-learning-in-python\u002F)\n\n[Understanding Machine Learning with Python | Pluralsight ](https:\u002F\u002Fwww.pluralsight.com\u002Fcourses\u002Fpython-understanding-machine-learning)\n\n[How to Think About Machine Learning Algorithms | Pluralsight](https:\u002F\u002Fwww.pluralsight.com\u002Fcourses\u002Fmachine-learning-algorithms)\n\n[Deep Learning Courses | Stanford Online](https:\u002F\u002Fonline.stanford.edu\u002Fcourses\u002Fcs230-deep-learning)\n\n[Deep Learning - UW Professional & Continuing Education](https:\u002F\u002Fwww.pce.uw.edu\u002Fcourses\u002Fdeep-learning)\n\n[Deep Learning Online Courses | Harvard University](https:\u002F\u002Fonline-learning.harvard.edu\u002Fcourse\u002Fdeep-learning-0)\n\n[Machine Learning for Everyone Courses | DataCamp](https:\u002F\u002Fwww.datacamp.com\u002Fcourses\u002Fintroduction-to-machine-learning-with-r)\n\n[Artificial Intelligence Expert Course: Platinum Edition | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fartificial-intelligence-exposed-future-10-extreme-edition\u002F)\n\n[Top Artificial Intelligence Courses Online | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=artificial%20intelligence)\n\n[Learn Artificial Intelligence with Online Courses and Lessons | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fartificial-intelligence)\n\n[Professional Certificate in Computer Science for Artificial Intelligence | edX](https:\u002F\u002Fwww.edx.org\u002Fprofessional-certificate\u002Fharvardx-computer-science-for-artifical-intelligence)\n\n[Artificial Intelligence Nanodegree program](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fai-artificial-intelligence-nanodegree--nd898)\n\n[Artificial Intelligence (AI) Online Courses | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fschool-of-ai)\n\n[Intro to Artificial Intelligence Course | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fintro-to-artificial-intelligence--cs271)\n\n[Edge AI for IoT Developers Course | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fintel-edge-ai-for-iot-developers-nanodegree--nd131)\n\n[Reasoning: Goal Trees and Rule-Based Expert Systems | MIT OpenCourseWare](https:\u002F\u002Focw.mit.edu\u002Fcourses\u002Felectrical-engineering-and-computer-science\u002F6-034-artificial-intelligence-fall-2010\u002Flecture-videos\u002Flecture-3-reasoning-goal-trees-and-rule-based-expert-systems\u002F)\n\n[Expert Systems and Applied Artificial Intelligence](https:\u002F\u002Fwww.umsl.edu\u002F~joshik\u002Fmsis480\u002Fchapt11.htm)\n\n[Autonomous Systems - Microsoft AI](https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002Fai\u002Fautonomous-systems)\n\n[Introduction to Microsoft Project Bonsai](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Flearn\u002Fautonomous-systems\u002Fintro-to-project-bonsai\u002F)\n\n[Machine teaching with the Microsoft Autonomous Systems platform](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Farchitecture\u002Fsolution-ideas\u002Farticles\u002Fautonomous-systems)\n\n[Autonomous Maritime Systems Training | AMC Search](https:\u002F\u002Fwww.amcsearch.com.au\u002Fams-training)\n\n[Top Autonomous Cars Courses Online | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fautonomous-cars\u002F)\n\n[Applied Control Systems 1: autonomous cars: Math + PID + MPC | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fapplied-systems-control-for-engineers-modelling-pid-mpc\u002F)\n\n[Learn Autonomous Robotics with Online Courses and Lessons | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fautonomous-robotics)\n\n[Artificial Intelligence Nanodegree program](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fai-artificial-intelligence-nanodegree--nd898)\n\n[Autonomous Systems Online Courses & Programs | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fschool-of-autonomous-systems)\n\n[Edge AI for IoT Developers Course | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fintel-edge-ai-for-iot-developers-nanodegree--nd131)\n\n[Autonomous Systems MOOC and Free Online Courses | MOOC List](https:\u002F\u002Fwww.mooc-list.com\u002Ftags\u002Fautonomous-systems)\n\n[Robotics and Autonomous Systems Graduate Program | Standford Online](https:\u002F\u002Fonline.stanford.edu\u002Fprograms\u002Frobotics-and-autonomous-systems-graduate-program)\n\n[Mobile Autonomous Systems Laboratory | MIT OpenCourseWare](https:\u002F\u002Focw.mit.edu\u002Fcourses\u002Felectrical-engineering-and-computer-science\u002F6-186-mobile-autonomous-systems-laboratory-january-iap-2005\u002Flecture-notes\u002F)\n\n## Deep Learning Tools, Libraries, and Frameworks\n\n[NVIDIA cuDNN](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcudnn) is a GPU-accelerated library of primitives for [deep neural networks](https:\u002F\u002Fdeveloper.nvidia.com\u002Fdeep-learning). cuDNN provides highly tuned implementations for standard routines such as forward and backward convolution, pooling, normalization, and activation layers. cuDNN accelerates widely used deep learning frameworks, including [Caffe2](https:\u002F\u002Fcaffe2.ai\u002F), [Chainer](https:\u002F\u002Fchainer.org\u002F), [Keras](https:\u002F\u002Fkeras.io\u002F), [MATLAB](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fdeep-learning.html), [MxNet](https:\u002F\u002Fmxnet.incubator.apache.org\u002F), [PyTorch](https:\u002F\u002Fpytorch.org\u002F), and [TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002F).\n\n[NVIDIA DLSS (Deep Learning Super Sampling)](https:\u002F\u002Fdeveloper.nvidia.com\u002Fdlss) is a temporal image upscaling AI rendering technology that increases graphics performance using dedicated Tensor Core AI processors on GeForce RTX™ GPUs. DLSS uses the power of a deep learning neural network to boost frame rates and generate beautiful, sharp images for your games.\n\n[AMD FidelityFX Super Resolution (FSR)](https:\u002F\u002Fwww.amd.com\u002Fen\u002Ftechnologies\u002Fradeon-software-fidelityfx) is an open source, high-quality solution for producing high resolution frames from lower resolution inputs. It uses a collection of cutting-edge Deep Learning algorithms with a particular emphasis on creating high-quality edges, giving large performance improvements compared to rendering at native resolution directly. FSR enables “practical performance” for costly render operations, such as hardware ray tracing for the AMD RDNA™ and AMD RDNA™ 2 architectures.\n\n[Intel Xe Super Sampling (XeSS)](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=Y9hfpf-SqEg) is a temporal image upscaling AI rendering technology that increases graphics performance similar to NVIDIA's [DLSS (Deep Learning Super Sampling)](https:\u002F\u002Fdeveloper.nvidia.com\u002Fdlss). Intel's Arc GPU architecture (early 2022) will have GPUs that feature dedicated Xe-cores to run XeSS. The GPUs will have Xe Matrix eXtenstions matrix (XMX) engines for hardware-accelerated AI processing. XeSS will be able to run on devices without XMX, including integrated graphics, though, the performance of XeSS will be lower on non-Intel graphics cards because it will be powered by [DP4a instruction](https:\u002F\u002Fwww.intel.com\u002Fcontent\u002Fdam\u002Fwww\u002Fpublic\u002Fus\u002Fen\u002Fdocuments\u002Freference-guides\u002F11th-gen-quick-reference-guide.pdf).\n\n[Jupyter Notebook](https:\u002F\u002Fjupyter.org\u002F) is an open-source web application that allows you to create and share documents that contain live code, equations, visualizations and narrative text. Jupyter is used widely in industries that do data cleaning and transformation, numerical simulation, statistical modeling, data visualization, data science, and machine learning.\n\n[Apache Spark](https:\u002F\u002Fspark.apache.org\u002F) is a unified analytics engine for large-scale data processing. It provides high-level APIs in Scala, Java, Python, and R, and an optimized engine that supports general computation graphs for data analysis. It also supports a rich set of higher-level tools including Spark SQL for SQL and DataFrames, MLlib for machine learning, GraphX for graph processing, and Structured Streaming for stream processing.\n\n[Apache Spark Connector for SQL Server and Azure SQL](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fsql-spark-connector) is a high-performance connector that enables you to use transactional data in big data analytics and persists results for ad-hoc queries or reporting. The connector allows you to use any SQL database, on-premises or in the cloud, as an input data source or output data sink for Spark jobs.\n\n[Apache PredictionIO](https:\u002F\u002Fpredictionio.apache.org\u002F) is an open source machine learning framework for developers, data scientists, and end users. It supports event collection, deployment of algorithms, evaluation, querying predictive results via REST APIs. It is based on scalable open source services like Hadoop, HBase (and other DBs), Elasticsearch, Spark and implements what is called a Lambda Architecture.\n\n[Cluster Manager for Apache Kafka(CMAK)](https:\u002F\u002Fgithub.com\u002Fyahoo\u002FCMAK) is a tool for managing [Apache Kafka](https:\u002F\u002Fkafka.apache.org\u002F) clusters.\n\n[BigDL](https:\u002F\u002Fbigdl-project.github.io\u002F) is a distributed deep learning library for Apache Spark. With BigDL, users can write their deep learning applications as standard Spark programs, which can directly run on top of existing Spark or Hadoop clusters.\n\n[Eclipse Deeplearning4J (DL4J)](https:\u002F\u002Fdeeplearning4j.konduit.ai\u002F) is a set of projects intended to support all the needs of a JVM-based(Scala, Kotlin, Clojure, and Groovy) deep learning application. This means starting with the raw data, loading and preprocessing it from wherever and whatever format it is in to building and tuning a wide variety of simple and complex deep learning networks.\n\n[Deep Learning Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdeep-learning.html) is a tool that provides a framework for designing and implementing deep neural networks with algorithms, pretrained models, and apps. You can use convolutional neural networks (ConvNets, CNNs) and long short-term memory (LSTM) networks to perform classification and regression on image, time-series, and text data. You can build network architectures such as generative adversarial networks (GANs) and Siamese networks using automatic differentiation, custom training loops, and shared weights. With the Deep Network Designer app, you can design, analyze, and train networks graphically. It can exchange models with TensorFlow™ and PyTorch through the ONNX format and import models from TensorFlow-Keras and Caffe. The toolbox supports transfer learning with DarkNet-53, ResNet-50, NASNet, SqueezeNet and many other pretrained models.\n\n[Reinforcement Learning Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Freinforcement-learning.html) is a tool that provides an app, functions, and a Simulink® block for training policies using reinforcement learning algorithms, including DQN, PPO, SAC, and DDPG. You can use these policies to implement controllers and decision-making algorithms for complex applications such as resource allocation, robotics, and autonomous systems.\n\n[Deep Learning HDL Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdeep-learning-hdl.html) is a tool that provides functions and tools to prototype and implement deep learning networks on FPGAs and SoCs. It provides pre-built bitstreams for running a variety of deep learning networks on supported Xilinx® and Intel® FPGA and SoC devices. Profiling and estimation tools let you customize a deep learning network by exploring design, performance, and resource utilization tradeoffs.\n\n[Parallel Computing Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmatlab-parallel-server.html) is a tool that lets you solve computationally and data-intensive problems using multicore processors, GPUs, and computer clusters. High-level constructs such as parallel for-loops, special array types, and parallelized numerical algorithms enable you to parallelize MATLAB® applications without CUDA or MPI programming. The toolbox lets you use parallel-enabled functions in MATLAB and other toolboxes. You can use the toolbox with Simulink® to run multiple simulations of a model in parallel. Programs and models can run in both interactive and batch modes.\n\n[XGBoost](https:\u002F\u002Fxgboost.readthedocs.io\u002F) is an optimized distributed gradient boosting library designed to be highly efficient, flexible and portable. It implements machine learning algorithms under the Gradient Boosting framework. XGBoost provides a parallel tree boosting (also known as GBDT, GBM) that solve many data science problems in a fast and accurate way. It supports distributed training on multiple machines, including AWS, GCE, Azure, and Yarn clusters. Also, it can be integrated with Flink, Spark and other cloud dataflow systems.\n\n[LIBSVM](https:\u002F\u002Fwww.csie.ntu.edu.tw\u002F~cjlin\u002Flibsvm\u002F) is an integrated software for support vector classification, (C-SVC, nu-SVC), regression (epsilon-SVR, nu-SVR) and distribution estimation (one-class SVM). It supports multi-class classification.\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) is a simple and efficient tool for data mining and data analysis. It is built on NumPy,SciPy, and mathplotlib.\n\n[TensorFlow](https:\u002F\u002Fwww.tensorflow.org) is an end-to-end open source platform for machine learning. It has a comprehensive, flexible ecosystem of tools, libraries and community resources that lets researchers push the state-of-the-art in ML and developers easily build and deploy ML powered applications.\n\n[Keras](https:\u002F\u002Fkeras.io) is a high-level neural networks API, written in Python and capable of running on top of TensorFlow, CNTK, or Theano.It was developed with a focus on enabling fast experimentation. It is capable of running on top of TensorFlow, Microsoft Cognitive Toolkit, R, Theano, or PlaidML.\n\n[PyTorch](https:\u002F\u002Fpytorch.org) is a library for deep learning on irregular input data such as graphs, point clouds, and manifolds. Primarily developed by Facebook's AI Research lab.\n\n[Azure Databricks](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fdatabricks\u002F) is a fast and collaborative Apache Spark-based big data analytics service designed for data science and data engineering. Azure Databricks, sets up your Apache Spark environment in minutes, autoscale, and collaborate on shared projects in an interactive workspace. Azure Databricks supports Python, Scala, R, Java, and SQL, as well as data science frameworks and libraries including TensorFlow, PyTorch, and scikit-learn.\n\n[Microsoft Cognitive Toolkit (CNTK)](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fcognitive-toolkit\u002F) is an open-source toolkit for commercial-grade distributed deep learning. It describes neural networks as a series of computational steps via a directed graph. CNTK allows the user to easily realize and combine popular model types such as feed-forward DNNs, convolutional neural networks (CNNs) and recurrent neural networks (RNNs\u002FLSTMs). CNTK implements stochastic gradient descent (SGD, error backpropagation) learning with automatic differentiation and parallelization across multiple GPUs and servers.\n\n[Tensorflow_macOS](https:\u002F\u002Fgithub.com\u002Fapple\u002Ftensorflow_macos) is a Mac-optimized version of TensorFlow and TensorFlow Addons for macOS 11.0+ accelerated using Apple's ML Compute framework.\n\n[Apache Airflow](https:\u002F\u002Fairflow.apache.org) is an open-source workflow management platform created by the community to programmatically author, schedule and monitor workflows. Install. Principles. Scalable. Airflow has a modular architecture and uses a message queue to orchestrate an arbitrary number of workers. Airflow is ready to scale to infinity.\n\n[Open Neural Network Exchange(ONNX)](https:\u002F\u002Fgithub.com\u002Fonnx) is an open ecosystem that empowers AI developers to choose the right tools as their project evolves. ONNX provides an open source format for AI models, both deep learning and traditional ML. It defines an extensible computation graph model, as well as definitions of built-in operators and standard data types.\n\n[Apache MXNet](https:\u002F\u002Fmxnet.apache.org\u002F) is a deep learning framework designed for both efficiency and flexibility. It allows you to mix symbolic and imperative programming to maximize efficiency and productivity. At its core, MXNet contains a dynamic dependency scheduler that automatically parallelizes both symbolic and imperative operations on the fly. A graph optimization layer on top of that makes symbolic execution fast and memory efficient. MXNet is portable and lightweight, scaling effectively to multiple GPUs and multiple machines. Support for Python, R, Julia, Scala, Go, Javascript and more.\n\n[AutoGluon](https:\u002F\u002Fautogluon.mxnet.io\u002Findex.html) is toolkit for Deep learning that automates machine learning tasks enabling you to easily achieve strong predictive performance in your applications. With just a few lines of code, you can train and deploy high-accuracy deep learning models on tabular, image, and text data.\n\n[Anaconda](https:\u002F\u002Fwww.anaconda.com\u002F) is a very popular Data Science platform for machine learning and deep learning that enables users to develop models, train them, and deploy them.\n\n[PlaidML](https:\u002F\u002Fgithub.com\u002Fplaidml\u002Fplaidml) is an advanced and portable tensor compiler for enabling deep learning on laptops, embedded devices, or other devices where the available computing hardware is not well supported or the available software stack contains unpalatable license restrictions.\n\n[OpenCV](https:\u002F\u002Fopencv.org) is a highly optimized library with focus on real-time computer vision applications. The C++, Python, and Java interfaces support Linux, MacOS, Windows, iOS, and Android.\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) is a Python module for machine learning built on top of SciPy, NumPy, and matplotlib, making it easier to apply robust and simple implementations of many popular machine learning algorithms.\n\n[Weka](https:\u002F\u002Fwww.cs.waikato.ac.nz\u002Fml\u002Fweka\u002F) is an open source machine learning software that can be accessed through a graphical user interface, standard terminal applications, or a Java API. It is widely used for teaching, research, and industrial applications, contains a plethora of built-in tools for standard machine learning tasks, and additionally gives transparent access to well-known toolboxes such as scikit-learn, R, and Deeplearning4j.\n\n[Caffe](https:\u002F\u002Fgithub.com\u002FBVLC\u002Fcaffe) is a deep learning framework made with expression, speed, and modularity in mind. It is developed by Berkeley AI Research (BAIR)\u002FThe Berkeley Vision and Learning Center (BVLC) and community contributors.\n\n[Theano](https:\u002F\u002Fgithub.com\u002FTheano\u002FTheano) is a Python library that allows you to define, optimize, and evaluate mathematical expressions involving multi-dimensional arrays efficiently including tight integration with NumPy.\n\n[Microsoft Project Bonsai](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fproject-bonsai\u002F) is a low-code AI platform that speeds AI-powered automation development and part of the Autonomous Systems suite from Microsoft. Bonsai is used to build AI components that can provide operator guidance or make independent decisions to optimize process variables, improve production efficiency, and reduce downtime.\n\n[Microsoft AirSim](https:\u002F\u002Fmicrosoft.github.io\u002FAirSim\u002Flidar.html) is a simulator for drones, cars and more, built on Unreal Engine (with an experimental Unity release). AirSim is open-source, cross platform, and supports [software-in-the-loop simulation](https:\u002F\u002Fwww.mathworks.com\u002Fhelp\u002F\u002F\u002Fecoder\u002Fsoftware-in-the-loop-sil-simulation.html) with popular flight controllers such as PX4 & ArduPilot and [hardware-in-loop](https:\u002F\u002Fwww.ni.com\u002Fen-us\u002Finnovations\u002Fwhite-papers\u002F17\u002Fwhat-is-hardware-in-the-loop-.html) with PX4 for physically and visually realistic simulations. It is developed as an Unreal plugin that can simply be dropped into any Unreal environment. AirSim is being developed  as a platform for AI research to experiment with deep learning, computer vision and reinforcement learning algorithms for autonomous vehicles.\n\n[CARLA](https:\u002F\u002Fgithub.com\u002Fcarla-simulator\u002Fcarla) is an open-source simulator for autonomous driving research. CARLA has been developed from the ground up to support development, training, and validation of autonomous driving systems. In addition to open-source code and protocols, CARLA provides open digital assets (urban layouts, buildings, vehicles) that were created for this purpose and can be used freely.\n\n[ROS\u002FROS2 bridge for CARLA(package)](https:\u002F\u002Fgithub.com\u002Fcarla-simulator\u002Fros-bridge) is a bridge that enables two-way communication between ROS and CARLA. The information from the CARLA server is translated to ROS topics. In the same way, the messages sent between nodes in ROS get translated to commands to be applied in CARLA.\n\n[ROS Toolbox](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fros.html) is a tool that provides an interface connecting MATLAB® and Simulink® with the Robot Operating System (ROS and ROS 2), enabling you to create a network of ROS nodes. The toolbox includes MATLAB functions and Simulink blocks to import, analyze, and play back ROS data recorded in rosbag files. You can also connect to a live ROS network to access ROS messages.\n\n[Robotics Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Frobotics.html) provides a toolbox that brings robotics specific functionality(designing, simulating, and testing manipulators, mobile robots, and humanoid robots) to MATLAB, exploiting the native capabilities of MATLAB (linear algebra, portability, graphics). The toolbox also supports mobile robots with functions for robot motion models (bicycle), path planning algorithms (bug, distance transform, D*, PRM), kinodynamic planning (lattice, RRT), localization (EKF, particle filter), map building (EKF) and simultaneous localization and mapping (EKF), and a Simulink model a of non-holonomic vehicle. The Toolbox also including a detailed Simulink model for a quadrotor flying robot.\n\n[Image Processing Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fimage.html) is a tool that provides a comprehensive set of reference-standard algorithms and workflow apps for image processing, analysis, visualization, and algorithm development. You can perform image segmentation, image enhancement, noise reduction, geometric transformations, image registration, and 3D image processing.\n\n[Computer Vision Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fcomputer-vision.html) is a tool that provides algorithms, functions, and apps for designing and testing computer vision, 3D vision, and video processing systems. You can perform object detection and tracking, as well as feature detection, extraction, and matching. You can automate calibration workflows for single, stereo, and fisheye cameras. For 3D vision, the toolbox supports visual and point cloud SLAM, stereo vision, structure from motion, and point cloud processing.\n\n[Robotics Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Frobotics.html) is a tool that provides a toolbox that brings robotics specific functionality(designing, simulating, and testing manipulators, mobile robots, and humanoid robots) to MATLAB, exploiting the native capabilities of MATLAB (linear algebra, portability, graphics). The toolbox also supports mobile robots with functions for robot motion models (bicycle), path planning algorithms (bug, distance transform, D*, PRM), kinodynamic planning (lattice, RRT), localization (EKF, particle filter), map building (EKF) and simultaneous localization and mapping (EKF), and a Simulink model a of non-holonomic vehicle. The Toolbox also including a detailed Simulink model for a quadrotor flying robot.\n\n[Model Predictive Control Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmodel-predictive-control.html) is a tool that provides functions, an app, and Simulink® blocks for designing and simulating controllers using linear and nonlinear model predictive control (MPC). The toolbox lets you specify plant and disturbance models, horizons, constraints, and weights. By running closed-loop simulations, you can evaluate controller performance.\n\n[Predictive Maintenance Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fpredictive-maintenance.html)  is a tool that lets you manage sensor data, design condition indicators, and estimate the remaining useful life (RUL) of a machine. The toolbox provides functions and an interactive app for exploring, extracting, and ranking features using data-based and model-based techniques, including statistical, spectral, and time-series analysis.\n\n[Vision HDL Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fvision-hdl.html) is a tool that provides pixel-streaming algorithms for the design and implementation of vision systems on FPGAs and ASICs. It provides a design framework that supports a diverse set of interface types, frame sizes, and frame rates. The image processing, video, and computer vision algorithms in the toolbox use an architecture appropriate for HDL implementations.\n\n[Automated Driving Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fautomated-driving.html) is a MATLAB tool that provides algorithms and tools for designing, simulating, and testing ADAS and autonomous driving systems. You can design and test vision and lidar perception systems, as well as sensor fusion, path planning, and vehicle controllers. Visualization tools include a bird’s-eye-view plot and scope for sensor coverage, detections and tracks, and displays for video, lidar, and maps. The toolbox lets you import and work with HERE HD Live Map data and OpenDRIVE® road networks. It also provides reference application examples for common ADAS and automated driving features, including FCW, AEB, ACC, LKA, and parking valet. The toolbox supports C\u002FC++ code generation for rapid prototyping and HIL testing, with support for sensor fusion, tracking, path planning, and vehicle controller algorithms.\n\n[UAV Toolbox](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fuav.html) is an application that provides tools and reference applications for designing, simulating, testing, and deploying unmanned aerial vehicle (UAV) and drone applications. You can design autonomous flight algorithms, UAV missions, and flight controllers. The Flight Log Analyzer app lets you interactively analyze 3D flight paths, telemetry information, and sensor readings from common flight log formats.\n\n[Navigation Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fnavigation.html) is a tool that provides algorithms and analysis tools for motion planning, simultaneous localization and mapping (SLAM), and inertial navigation. The toolbox includes customizable search and sampling-based path planners, as well as metrics for validating and comparing paths. You can create 2D and 3D map representations, generate maps using SLAM algorithms, and interactively visualize and debug map generation with the SLAM map builder app.\n\n[Lidar Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Flidar.html) is a tool that provides algorithms, functions, and apps for designing, analyzing, and testing lidar processing systems. You can perform object detection and tracking, semantic segmentation, shape fitting, lidar registration, and obstacle detection. Lidar Toolbox supports lidar-camera cross calibration for workflows that combine computer vision and lidar processing.\n\n[Mapping Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmapping.html) is a tool that provides algorithms and functions for transforming geographic data and creating map displays. You can visualize your data in a geographic context, build map displays from more than 60 map projections, and transform data from a variety of sources into a consistent geographic coordinate system.\n\n\n# Reinforcement Learning Development\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_d7c013b9339b.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## Reinforcement Learning Learning Resources\n\n[Reinforcement Learning](https:\u002F\u002Fwww.ibm.com\u002Fcloud\u002Flearn\u002Fdeep-learning#toc-deep-learn-md_Q_Of3) is a subset of machine learning, which is a neural network with three or more layers. These neural networks attempt to simulate the behavior of the human brain,though, far from matching its ability. This allows the neural networks to \"learn\" from a process in which a model learns to become more accurate for performing an action in an environment based on feedback in order to maximize the reward. The Learning can be [supervised](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FSupervised_learning), [semi-supervised](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FSemi-supervised_learning) or [unsupervised](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FUnsupervised_learning).\n\n[Top Reinforcement Learning Courses | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=reinforcement%20learning)\n\n[Top Reinforcement Learning Courses | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Freinforcement-learning\u002F)\n\n[Top Reinforcement Learning Courses | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Freinforcement-learning--ud600)\n\n[Reinforcement Learning Courses | Stanford Online](https:\u002F\u002Fonline.stanford.edu\u002Fcourses\u002Fxcs234-reinforcement-learning)\n\n[Deep Learning Online Courses | NVIDIA](https:\u002F\u002Fwww.nvidia.com\u002Fen-us\u002Ftraining\u002Fonline\u002F)\n\n[Top Deep Learning Courses Online | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=deep%20learning)\n\n[Top Deep Learning Courses Online | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fdeep-learning\u002F)\n\n[Learn Deep Learning with Online Courses and Lessons | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fdeep-learning)\n\n[Deep Learning Online Course Nanodegree | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fdeep-learning-nanodegree--nd101)\n\n[Machine Learning Course by Andrew Ng | Coursera](https:\u002F\u002Fwww.coursera.org\u002Flearn\u002Fmachine-learning?)\n\n[Machine Learning Engineering for Production (MLOps) course by Andrew Ng | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fspecializations\u002Fmachine-learning-engineering-for-production-mlops)\n\n[Data Science: Deep Learning and Neural Networks in Python | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fdata-science-deep-learning-in-python\u002F)\n\n[Understanding Machine Learning with Python | Pluralsight ](https:\u002F\u002Fwww.pluralsight.com\u002Fcourses\u002Fpython-understanding-machine-learning)\n\n[How to Think About Machine Learning Algorithms | Pluralsight](https:\u002F\u002Fwww.pluralsight.com\u002Fcourses\u002Fmachine-learning-algorithms)\n\n[Deep Learning Courses | Stanford Online](https:\u002F\u002Fonline.stanford.edu\u002Fcourses\u002Fcs230-deep-learning)\n\n[Deep Learning - UW Professional & Continuing Education](https:\u002F\u002Fwww.pce.uw.edu\u002Fcourses\u002Fdeep-learning)\n\n[Deep Learning Online Courses | Harvard University](https:\u002F\u002Fonline-learning.harvard.edu\u002Fcourse\u002Fdeep-learning-0)\n\n[Machine Learning for Everyone Courses | DataCamp](https:\u002F\u002Fwww.datacamp.com\u002Fcourses\u002Fintroduction-to-machine-learning-with-r)\n\n[Artificial Intelligence Expert Course: Platinum Edition | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fartificial-intelligence-exposed-future-10-extreme-edition\u002F)\n\n[Top Artificial Intelligence Courses Online | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=artificial%20intelligence)\n\n[Learn Artificial Intelligence with Online Courses and Lessons | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fartificial-intelligence)\n\n[Professional Certificate in Computer Science for Artificial Intelligence | edX](https:\u002F\u002Fwww.edx.org\u002Fprofessional-certificate\u002Fharvardx-computer-science-for-artifical-intelligence)\n\n[Artificial Intelligence Nanodegree program](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fai-artificial-intelligence-nanodegree--nd898)\n\n[Artificial Intelligence (AI) Online Courses | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fschool-of-ai)\n\n[Intro to Artificial Intelligence Course | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fintro-to-artificial-intelligence--cs271)\n\n[Edge AI for IoT Developers Course | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fintel-edge-ai-for-iot-developers-nanodegree--nd131)\n\n[Reasoning: Goal Trees and Rule-Based Expert Systems | MIT OpenCourseWare](https:\u002F\u002Focw.mit.edu\u002Fcourses\u002Felectrical-engineering-and-computer-science\u002F6-034-artificial-intelligence-fall-2010\u002Flecture-videos\u002Flecture-3-reasoning-goal-trees-and-rule-based-expert-systems\u002F)\n\n[Expert Systems and Applied Artificial Intelligence](https:\u002F\u002Fwww.umsl.edu\u002F~joshik\u002Fmsis480\u002Fchapt11.htm)\n\n[Autonomous Systems - Microsoft AI](https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002Fai\u002Fautonomous-systems)\n\n[Introduction to Microsoft Project Bonsai](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Flearn\u002Fautonomous-systems\u002Fintro-to-project-bonsai\u002F)\n\n[Machine teaching with the Microsoft Autonomous Systems platform](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Farchitecture\u002Fsolution-ideas\u002Farticles\u002Fautonomous-systems)\n\n[Autonomous Maritime Systems Training | AMC Search](https:\u002F\u002Fwww.amcsearch.com.au\u002Fams-training)\n\n[Top Autonomous Cars Courses Online | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fautonomous-cars\u002F)\n\n[Applied Control Systems 1: autonomous cars: Math + PID + MPC | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fapplied-systems-control-for-engineers-modelling-pid-mpc\u002F)\n\n[Learn Autonomous Robotics with Online Courses and Lessons | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fautonomous-robotics)\n\n[Artificial Intelligence Nanodegree program](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fai-artificial-intelligence-nanodegree--nd898)\n\n[Autonomous Systems Online Courses & Programs | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fschool-of-autonomous-systems)\n\n[Edge AI for IoT Developers Course | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fintel-edge-ai-for-iot-developers-nanodegree--nd131)\n\n[Autonomous Systems MOOC and Free Online Courses | MOOC List](https:\u002F\u002Fwww.mooc-list.com\u002Ftags\u002Fautonomous-systems)\n\n[Robotics and Autonomous Systems Graduate Program | Standford Online](https:\u002F\u002Fonline.stanford.edu\u002Fprograms\u002Frobotics-and-autonomous-systems-graduate-program)\n\n[Mobile Autonomous Systems Laboratory | MIT OpenCourseWare](https:\u002F\u002Focw.mit.edu\u002Fcourses\u002Felectrical-engineering-and-computer-science\u002F6-186-mobile-autonomous-systems-laboratory-january-iap-2005\u002Flecture-notes\u002F)\n\n## Reinforcement Learning Tools, Libraries, and Frameworks\n\n[OpenAI](https:\u002F\u002Fgym.openai.com\u002F) is an open source Python library for developing and comparing reinforcement learning algorithms by providing a standard API to communicate between learning algorithms and environments, as well as a standard set of environments compliant with that API.\n\n[ReinforcementLearning.jl](https:\u002F\u002Fjuliareinforcementlearning.org\u002F) is a collection of tools for doing reinforcement learning research in Julia.\n\n[Reinforcement Learning Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Freinforcement-learning.html) is a tool that provides an app, functions, and a Simulink® block for training policies using reinforcement learning algorithms, including DQN, PPO, SAC, and DDPG. You can use these policies to implement controllers and decision-making algorithms for complex applications such as resource allocation, robotics, and autonomous systems.\n\n[Amazon SageMaker](https:\u002F\u002Faws.amazon.com\u002Frobomaker\u002F) is a fully managed service that provides every developer and data scientist with the ability to build, train, and deploy machine learning (ML) models quickly.\n\n[AWS RoboMaker](https:\u002F\u002Faws.amazon.com\u002Frobomaker\u002F) is a service that provides a fully-managed, scalable infrastructure for simulation that customers use for multi-robot simulation and CI\u002FCD integration with regression testing in simulation.\n\n[TensorFlow](https:\u002F\u002Fwww.tensorflow.org) is an end-to-end open source platform for machine learning. It has a comprehensive, flexible ecosystem of tools, libraries and community resources that lets researchers push the state-of-the-art in ML and developers easily build and deploy ML powered applications.\n\n[Keras](https:\u002F\u002Fkeras.io) is a high-level neural networks API, written in Python and capable of running on top of TensorFlow, CNTK, or Theano.It was developed with a focus on enabling fast experimentation. It is capable of running on top of TensorFlow, Microsoft Cognitive Toolkit, R, Theano, or PlaidML.\n\n[PyTorch](https:\u002F\u002Fpytorch.org) is a library for deep learning on irregular input data such as graphs, point clouds, and manifolds. Primarily developed by Facebook's AI Research lab.\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) is a simple and efficient tool for data mining and data analysis. It is built on NumPy,SciPy, and mathplotlib.\n\n[NVIDIA cuDNN](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcudnn) is a GPU-accelerated library of primitives for [deep neural networks](https:\u002F\u002Fdeveloper.nvidia.com\u002Fdeep-learning). cuDNN provides highly tuned implementations for standard routines such as forward and backward convolution, pooling, normalization, and activation layers. cuDNN accelerates widely used deep learning frameworks, including [Caffe2](https:\u002F\u002Fcaffe2.ai\u002F), [Chainer](https:\u002F\u002Fchainer.org\u002F), [Keras](https:\u002F\u002Fkeras.io\u002F), [MATLAB](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fdeep-learning.html), [MxNet](https:\u002F\u002Fmxnet.incubator.apache.org\u002F), [PyTorch](https:\u002F\u002Fpytorch.org\u002F), and [TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002F).\n\n[Jupyter Notebook](https:\u002F\u002Fjupyter.org\u002F) is an open-source web application that allows you to create and share documents that contain live code, equations, visualizations and narrative text. Jupyter is used widely in industries that do data cleaning and transformation, numerical simulation, statistical modeling, data visualization, data science, and machine learning.\n\n[Apache Spark](https:\u002F\u002Fspark.apache.org\u002F) is a unified analytics engine for large-scale data processing. It provides high-level APIs in Scala, Java, Python, and R, and an optimized engine that supports general computation graphs for data analysis. It also supports a rich set of higher-level tools including Spark SQL for SQL and DataFrames, MLlib for machine learning, GraphX for graph processing, and Structured Streaming for stream processing.\n\n[Apache Spark Connector for SQL Server and Azure SQL](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fsql-spark-connector) is a high-performance connector that enables you to use transactional data in big data analytics and persists results for ad-hoc queries or reporting. The connector allows you to use any SQL database, on-premises or in the cloud, as an input data source or output data sink for Spark jobs.\n\n[Apache PredictionIO](https:\u002F\u002Fpredictionio.apache.org\u002F) is an open source machine learning framework for developers, data scientists, and end users. It supports event collection, deployment of algorithms, evaluation, querying predictive results via REST APIs. It is based on scalable open source services like Hadoop, HBase (and other DBs), Elasticsearch, Spark and implements what is called a Lambda Architecture.\n\n[Cluster Manager for Apache Kafka(CMAK)](https:\u002F\u002Fgithub.com\u002Fyahoo\u002FCMAK) is a tool for managing [Apache Kafka](https:\u002F\u002Fkafka.apache.org\u002F) clusters.\n\n[BigDL](https:\u002F\u002Fbigdl-project.github.io\u002F) is a distributed deep learning library for Apache Spark. With BigDL, users can write their deep learning applications as standard Spark programs, which can directly run on top of existing Spark or Hadoop clusters.\n\n[Eclipse Deeplearning4J (DL4J)](https:\u002F\u002Fdeeplearning4j.konduit.ai\u002F) is a set of projects intended to support all the needs of a JVM-based(Scala, Kotlin, Clojure, and Groovy) deep learning application. This means starting with the raw data, loading and preprocessing it from wherever and whatever format it is in to building and tuning a wide variety of simple and complex deep learning networks.\n\n[Deep Learning Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdeep-learning.html) is a tool that provides a framework for designing and implementing deep neural networks with algorithms, pretrained models, and apps. You can use convolutional neural networks (ConvNets, CNNs) and long short-term memory (LSTM) networks to perform classification and regression on image, time-series, and text data. You can build network architectures such as generative adversarial networks (GANs) and Siamese networks using automatic differentiation, custom training loops, and shared weights. With the Deep Network Designer app, you can design, analyze, and train networks graphically. It can exchange models with TensorFlow™ and PyTorch through the ONNX format and import models from TensorFlow-Keras and Caffe. The toolbox supports transfer learning with DarkNet-53, ResNet-50, NASNet, SqueezeNet and many other pretrained models.\n\n[Deep Learning HDL Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdeep-learning-hdl.html) is a tool that provides functions and tools to prototype and implement deep learning networks on FPGAs and SoCs. It provides pre-built bitstreams for running a variety of deep learning networks on supported Xilinx® and Intel® FPGA and SoC devices. Profiling and estimation tools let you customize a deep learning network by exploring design, performance, and resource utilization tradeoffs.\n\n[Parallel Computing Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmatlab-parallel-server.html) is a tool that lets you solve computationally and data-intensive problems using multicore processors, GPUs, and computer clusters. High-level constructs such as parallel for-loops, special array types, and parallelized numerical algorithms enable you to parallelize MATLAB® applications without CUDA or MPI programming. The toolbox lets you use parallel-enabled functions in MATLAB and other toolboxes. You can use the toolbox with Simulink® to run multiple simulations of a model in parallel. Programs and models can run in both interactive and batch modes.\n\n[XGBoost](https:\u002F\u002Fxgboost.readthedocs.io\u002F) is an optimized distributed gradient boosting library designed to be highly efficient, flexible and portable. It implements machine learning algorithms under the Gradient Boosting framework. XGBoost provides a parallel tree boosting (also known as GBDT, GBM) that solve many data science problems in a fast and accurate way. It supports distributed training on multiple machines, including AWS, GCE, Azure, and Yarn clusters. Also, it can be integrated with Flink, Spark and other cloud dataflow systems.\n\n[LIBSVM](https:\u002F\u002Fwww.csie.ntu.edu.tw\u002F~cjlin\u002Flibsvm\u002F) is an integrated software for support vector classification, (C-SVC, nu-SVC), regression (epsilon-SVR, nu-SVR) and distribution estimation (one-class SVM). It supports multi-class classification.\n\n[Azure Databricks](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fdatabricks\u002F) is a fast and collaborative Apache Spark-based big data analytics service designed for data science and data engineering. Azure Databricks, sets up your Apache Spark environment in minutes, autoscale, and collaborate on shared projects in an interactive workspace. Azure Databricks supports Python, Scala, R, Java, and SQL, as well as data science frameworks and libraries including TensorFlow, PyTorch, and scikit-learn.\n\n[Microsoft Cognitive Toolkit (CNTK)](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fcognitive-toolkit\u002F) is an open-source toolkit for commercial-grade distributed deep learning. It describes neural networks as a series of computational steps via a directed graph. CNTK allows the user to easily realize and combine popular model types such as feed-forward DNNs, convolutional neural networks (CNNs) and recurrent neural networks (RNNs\u002FLSTMs). CNTK implements stochastic gradient descent (SGD, error backpropagation) learning with automatic differentiation and parallelization across multiple GPUs and servers.\n\n[Tensorflow_macOS](https:\u002F\u002Fgithub.com\u002Fapple\u002Ftensorflow_macos) is a Mac-optimized version of TensorFlow and TensorFlow Addons for macOS 11.0+ accelerated using Apple's ML Compute framework.\n\n[Apache Airflow](https:\u002F\u002Fairflow.apache.org) is an open-source workflow management platform created by the community to programmatically author, schedule and monitor workflows. Install. Principles. Scalable. Airflow has a modular architecture and uses a message queue to orchestrate an arbitrary number of workers. Airflow is ready to scale to infinity.\n\n[Open Neural Network Exchange(ONNX)](https:\u002F\u002Fgithub.com\u002Fonnx) is an open ecosystem that empowers AI developers to choose the right tools as their project evolves. ONNX provides an open source format for AI models, both deep learning and traditional ML. It defines an extensible computation graph model, as well as definitions of built-in operators and standard data types.\n\n[Apache MXNet](https:\u002F\u002Fmxnet.apache.org\u002F) is a deep learning framework designed for both efficiency and flexibility. It allows you to mix symbolic and imperative programming to maximize efficiency and productivity. At its core, MXNet contains a dynamic dependency scheduler that automatically parallelizes both symbolic and imperative operations on the fly. A graph optimization layer on top of that makes symbolic execution fast and memory efficient. MXNet is portable and lightweight, scaling effectively to multiple GPUs and multiple machines. Support for Python, R, Julia, Scala, Go, Javascript and more.\n\n[AutoGluon](https:\u002F\u002Fautogluon.mxnet.io\u002Findex.html) is toolkit for Deep learning that automates machine learning tasks enabling you to easily achieve strong predictive performance in your applications. With just a few lines of code, you can train and deploy high-accuracy deep learning models on tabular, image, and text data.\n\n[Anaconda](https:\u002F\u002Fwww.anaconda.com\u002F) is a very popular Data Science platform for machine learning and deep learning that enables users to develop models, train them, and deploy them.\n\n[PlaidML](https:\u002F\u002Fgithub.com\u002Fplaidml\u002Fplaidml) is an advanced and portable tensor compiler for enabling deep learning on laptops, embedded devices, or other devices where the available computing hardware is not well supported or the available software stack contains unpalatable license restrictions.\n\n[OpenCV](https:\u002F\u002Fopencv.org) is a highly optimized library with focus on real-time computer vision applications. The C++, Python, and Java interfaces support Linux, MacOS, Windows, iOS, and Android.\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) is a Python module for machine learning built on top of SciPy, NumPy, and matplotlib, making it easier to apply robust and simple implementations of many popular machine learning algorithms.\n\n[Weka](https:\u002F\u002Fwww.cs.waikato.ac.nz\u002Fml\u002Fweka\u002F) is an open source machine learning software that can be accessed through a graphical user interface, standard terminal applications, or a Java API. It is widely used for teaching, research, and industrial applications, contains a plethora of built-in tools for standard machine learning tasks, and additionally gives transparent access to well-known toolboxes such as scikit-learn, R, and Deeplearning4j.\n\n[Caffe](https:\u002F\u002Fgithub.com\u002FBVLC\u002Fcaffe) is a deep learning framework made with expression, speed, and modularity in mind. It is developed by Berkeley AI Research (BAIR)\u002FThe Berkeley Vision and Learning Center (BVLC) and community contributors.\n\n[Theano](https:\u002F\u002Fgithub.com\u002FTheano\u002FTheano) is a Python library that allows you to define, optimize, and evaluate mathematical expressions involving multi-dimensional arrays efficiently including tight integration with NumPy.\n\n[Microsoft Project Bonsai](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fproject-bonsai\u002F) is a low-code AI platform that speeds AI-powered automation development and part of the Autonomous Systems suite from Microsoft. Bonsai is used to build AI components that can provide operator guidance or make independent decisions to optimize process variables, improve production efficiency, and reduce downtime.\n\n[Microsoft AirSim](https:\u002F\u002Fmicrosoft.github.io\u002FAirSim\u002Flidar.html) is a simulator for drones, cars and more, built on Unreal Engine (with an experimental Unity release). AirSim is open-source, cross platform, and supports [software-in-the-loop simulation](https:\u002F\u002Fwww.mathworks.com\u002Fhelp\u002F\u002F\u002Fecoder\u002Fsoftware-in-the-loop-sil-simulation.html) with popular flight controllers such as PX4 & ArduPilot and [hardware-in-loop](https:\u002F\u002Fwww.ni.com\u002Fen-us\u002Finnovations\u002Fwhite-papers\u002F17\u002Fwhat-is-hardware-in-the-loop-.html) with PX4 for physically and visually realistic simulations. It is developed as an Unreal plugin that can simply be dropped into any Unreal environment. AirSim is being developed  as a platform for AI research to experiment with deep learning, computer vision and reinforcement learning algorithms for autonomous vehicles.\n\n[CARLA](https:\u002F\u002Fgithub.com\u002Fcarla-simulator\u002Fcarla) is an open-source simulator for autonomous driving research. CARLA has been developed from the ground up to support development, training, and validation of autonomous driving systems. In addition to open-source code and protocols, CARLA provides open digital assets (urban layouts, buildings, vehicles) that were created for this purpose and can be used freely.\n\n[ROS\u002FROS2 bridge for CARLA(package)](https:\u002F\u002Fgithub.com\u002Fcarla-simulator\u002Fros-bridge) is a bridge that enables two-way communication between ROS and CARLA. The information from the CARLA server is translated to ROS topics. In the same way, the messages sent between nodes in ROS get translated to commands to be applied in CARLA.\n\n[ROS Toolbox](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fros.html) is a tool that provides an interface connecting MATLAB® and Simulink® with the Robot Operating System (ROS and ROS 2), enabling you to create a network of ROS nodes. The toolbox includes MATLAB functions and Simulink blocks to import, analyze, and play back ROS data recorded in rosbag files. You can also connect to a live ROS network to access ROS messages.\n\n[Robotics Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Frobotics.html) provides a toolbox that brings robotics specific functionality(designing, simulating, and testing manipulators, mobile robots, and humanoid robots) to MATLAB, exploiting the native capabilities of MATLAB (linear algebra, portability, graphics). The toolbox also supports mobile robots with functions for robot motion models (bicycle), path planning algorithms (bug, distance transform, D*, PRM), kinodynamic planning (lattice, RRT), localization (EKF, particle filter), map building (EKF) and simultaneous localization and mapping (EKF), and a Simulink model a of non-holonomic vehicle. The Toolbox also including a detailed Simulink model for a quadrotor flying robot.\n\n[Image Processing Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fimage.html) is a tool that provides a comprehensive set of reference-standard algorithms and workflow apps for image processing, analysis, visualization, and algorithm development. You can perform image segmentation, image enhancement, noise reduction, geometric transformations, image registration, and 3D image processing.\n\n[Computer Vision Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fcomputer-vision.html) is a tool that provides algorithms, functions, and apps for designing and testing computer vision, 3D vision, and video processing systems. You can perform object detection and tracking, as well as feature detection, extraction, and matching. You can automate calibration workflows for single, stereo, and fisheye cameras. For 3D vision, the toolbox supports visual and point cloud SLAM, stereo vision, structure from motion, and point cloud processing.\n\n[Robotics Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Frobotics.html) is a tool that provides a toolbox that brings robotics specific functionality(designing, simulating, and testing manipulators, mobile robots, and humanoid robots) to MATLAB, exploiting the native capabilities of MATLAB (linear algebra, portability, graphics). The toolbox also supports mobile robots with functions for robot motion models (bicycle), path planning algorithms (bug, distance transform, D*, PRM), kinodynamic planning (lattice, RRT), localization (EKF, particle filter), map building (EKF) and simultaneous localization and mapping (EKF), and a Simulink model a of non-holonomic vehicle. The Toolbox also including a detailed Simulink model for a quadrotor flying robot.\n\n[Model Predictive Control Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmodel-predictive-control.html) is a tool that provides functions, an app, and Simulink® blocks for designing and simulating controllers using linear and nonlinear model predictive control (MPC). The toolbox lets you specify plant and disturbance models, horizons, constraints, and weights. By running closed-loop simulations, you can evaluate controller performance.\n\n[Predictive Maintenance Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fpredictive-maintenance.html)  is a tool that lets you manage sensor data, design condition indicators, and estimate the remaining useful life (RUL) of a machine. The toolbox provides functions and an interactive app for exploring, extracting, and ranking features using data-based and model-based techniques, including statistical, spectral, and time-series analysis.\n\n[Vision HDL Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fvision-hdl.html) is a tool that provides pixel-streaming algorithms for the design and implementation of vision systems on FPGAs and ASICs. It provides a design framework that supports a diverse set of interface types, frame sizes, and frame rates. The image processing, video, and computer vision algorithms in the toolbox use an architecture appropriate for HDL implementations.\n\n[Automated Driving Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fautomated-driving.html) is a MATLAB tool that provides algorithms and tools for designing, simulating, and testing ADAS and autonomous driving systems. You can design and test vision and lidar perception systems, as well as sensor fusion, path planning, and vehicle controllers. Visualization tools include a bird’s-eye-view plot and scope for sensor coverage, detections and tracks, and displays for video, lidar, and maps. The toolbox lets you import and work with HERE HD Live Map data and OpenDRIVE® road networks. It also provides reference application examples for common ADAS and automated driving features, including FCW, AEB, ACC, LKA, and parking valet. The toolbox supports C\u002FC++ code generation for rapid prototyping and HIL testing, with support for sensor fusion, tracking, path planning, and vehicle controller algorithms.\n\n[Navigation Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fnavigation.html) is a tool that provides algorithms and analysis tools for motion planning, simultaneous localization and mapping (SLAM), and inertial navigation. The toolbox includes customizable search and sampling-based path planners, as well as metrics for validating and comparing paths. You can create 2D and 3D map representations, generate maps using SLAM algorithms, and interactively visualize and debug map generation with the SLAM map builder app.\n\n[UAV Toolbox](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fuav.html) is an application that provides tools and reference applications for designing, simulating, testing, and deploying unmanned aerial vehicle (UAV) and drone applications. You can design autonomous flight algorithms, UAV missions, and flight controllers. The Flight Log Analyzer app lets you interactively analyze 3D flight paths, telemetry information, and sensor readings from common flight log formats.\n\n[Lidar Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Flidar.html) is a tool that provides algorithms, functions, and apps for designing, analyzing, and testing lidar processing systems. You can perform object detection and tracking, semantic segmentation, shape fitting, lidar registration, and obstacle detection. Lidar Toolbox supports lidar-camera cross calibration for workflows that combine computer vision and lidar processing.\n\n[Mapping Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmapping.html) is a tool that provides algorithms and functions for transforming geographic data and creating map displays. You can visualize your data in a geographic context, build map displays from more than 60 map projections, and transform data from a variety of sources into a consistent geographic coordinate system.\n\n\n# Computer Vision Development\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_01c04b532852.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## Computer Vision Learning Resources\n\n[Computer Vision](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Foverview\u002Fwhat-is-computer-vision\u002F) is a field of Artificial Intelligence (AI) that focuses on enabling computers to identify and understand objects and people in images and videos.\n\n[OpenCV Courses](https:\u002F\u002Fopencv.org\u002Fcourses\u002F)\n\n[Exploring Computer Vision in Microsoft Azure](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Flearn\u002Fpaths\u002Fexplore-computer-vision-microsoft-azure\u002F)\n\n[Top Computer Vision Courses Online | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?languages=en&query=computer%20vision)\n\n[Top Computer Vision Courses Online | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fcomputer-vision\u002F)\n\n[Learn Computer Vision with Online Courses and Lessons | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fcomputer-vision)\n\n[Computer Vision and Image Processing Fundamentals | edX](https:\u002F\u002Fwww.edx.org\u002Fcourse\u002Fcomputer-vision-and-image-processing-fundamentals)\n\n[Introduction to Computer Vision Courses | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fintroduction-to-computer-vision--ud810)\n\n[Computer Vision Nanodegree program | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fcomputer-vision-nanodegree--nd891)\n\n[Machine Vision Course |MIT Open Courseware ](https:\u002F\u002Focw.mit.edu\u002Fcourses\u002Felectrical-engineering-and-computer-science\u002F6-801-machine-vision-fall-2004\u002F)\n\n[Computer Vision Training Courses | NobleProg](https:\u002F\u002Fwww.nobleprog.com\u002Fcomputer-vision-training)\n\n[Visual Computing Graduate Program | Stanford Online](https:\u002F\u002Fonline.stanford.edu\u002Fprograms\u002Fvisual-computing-graduate-program)\n\n## Computer Vision Tools, Libraries, and Frameworks\n\n[OpenCV](https:\u002F\u002Fopencv.org) is a highly optimized library with focus on real-time computer vision applications. The C++, Python, and Java interfaces support Linux, MacOS, Windows, iOS, and Android.\n\n[Microsoft Computer Vision Recipes](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fcomputervision-recipes) is a project that provides examples and best practice guidelines for building computer vision systems. This allows people to build a comprehensive set of tools and examples that leverage recent advances in Computer Vision algorithms, neural architectures, and operationalizing such systems. Creatin from existing state-of-the-art libraries and build additional utility around loading image data, optimizing and evaluating models, and scaling up to the cloud. \n\n[Microsoft Cognitive Toolkit (CNTK)](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fcognitive-toolkit\u002F) is an open-source toolkit for commercial-grade distributed deep learning. It describes neural networks as a series of computational steps via a directed graph. CNTK allows the user to easily realize and combine popular model types such as feed-forward DNNs, convolutional neural networks (CNNs) and recurrent neural networks (RNNs\u002FLSTMs). CNTK implements stochastic gradient descent (SGD, error backpropagation) learning with automatic differentiation and parallelization across multiple GPUs and servers.\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) is a Python module for machine learning built on top of SciPy, NumPy, and matplotlib, making it easier to apply robust and simple implementations of many popular machine learning algorithms.\n\n[NVIDIA cuDNN](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcudnn) is a GPU-accelerated library of primitives for [deep neural networks](https:\u002F\u002Fdeveloper.nvidia.com\u002Fdeep-learning). cuDNN provides highly tuned implementations for standard routines such as forward and backward convolution, pooling, normalization, and activation layers. cuDNN accelerates widely used deep learning frameworks, including [Caffe2](https:\u002F\u002Fcaffe2.ai\u002F), [Chainer](https:\u002F\u002Fchainer.org\u002F), [Keras](https:\u002F\u002Fkeras.io\u002F), [MATLAB](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fdeep-learning.html), [MxNet](https:\u002F\u002Fmxnet.incubator.apache.org\u002F), [PyTorch](https:\u002F\u002Fpytorch.org\u002F), and [TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002F).\n\n[Automated Driving Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fautomated-driving.html) is a MATLAB tool that provides algorithms and tools for designing, simulating, and testing ADAS and autonomous driving systems. You can design and test vision and lidar perception systems, as well as sensor fusion, path planning, and vehicle controllers. Visualization tools include a bird’s-eye-view plot and scope for sensor coverage, detections and tracks, and displays for video, lidar, and maps. The toolbox lets you import and work with HERE HD Live Map data and OpenDRIVE® road networks. It also provides reference application examples for common ADAS and automated driving features, including FCW, AEB, ACC, LKA, and parking valet. The toolbox supports C\u002FC++ code generation for rapid prototyping and HIL testing, with support for sensor fusion, tracking, path planning, and vehicle controller algorithms.\n\n[LRSLibrary](https:\u002F\u002Fgithub.com\u002Fandrewssobral\u002Flrslibrary) is a Low-Rank and Sparse Tools for Background Modeling and Subtraction in Videos. The library was designed for moving object detection in videos, but it can be also used for other computer vision and machine learning problems.\n\n[Image Processing Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fimage.html) is a tool that provides a comprehensive set of reference-standard algorithms and workflow apps for image processing, analysis, visualization, and algorithm development. You can perform image segmentation, image enhancement, noise reduction, geometric transformations, image registration, and 3D image processing.\n\n[Computer Vision Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fcomputer-vision.html) is a tool that provides algorithms, functions, and apps for designing and testing computer vision, 3D vision, and video processing systems. You can perform object detection and tracking, as well as feature detection, extraction, and matching. You can automate calibration workflows for single, stereo, and fisheye cameras. For 3D vision, the toolbox supports visual and point cloud SLAM, stereo vision, structure from motion, and point cloud processing.\n\n[Statistics and Machine Learning Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fstatistics.html) is a tool that provides functions and apps to describe, analyze, and model data. You can use descriptive statistics, visualizations, and clustering for exploratory data analysis; fit probability distributions to data; generate random numbers for Monte Carlo simulations, and perform hypothesis tests. Regression and classification algorithms let you draw inferences from data and build predictive models either interactively, using the Classification and Regression Learner apps, or programmatically, using AutoML.\n\n[Lidar Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Flidar.html) is a tool that provides algorithms, functions, and apps for designing, analyzing, and testing lidar processing systems. You can perform object detection and tracking, semantic segmentation, shape fitting, lidar registration, and obstacle detection. Lidar Toolbox supports lidar-camera cross calibration for workflows that combine computer vision and lidar processing.\n\n[Mapping Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmapping.html) is a tool that provides algorithms and functions for transforming geographic data and creating map displays. You can visualize your data in a geographic context, build map displays from more than 60 map projections, and transform data from a variety of sources into a consistent geographic coordinate system.\n\n[UAV Toolbox](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fuav.html) is an application that provides tools and reference applications for designing, simulating, testing, and deploying unmanned aerial vehicle (UAV) and drone applications. You can design autonomous flight algorithms, UAV missions, and flight controllers. The Flight Log Analyzer app lets you interactively analyze 3D flight paths, telemetry information, and sensor readings from common flight log formats.\n\n[Parallel Computing Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmatlab-parallel-server.html) is a tool that lets you solve computationally and data-intensive problems using multicore processors, GPUs, and computer clusters. High-level constructs such as parallel for-loops, special array types, and parallelized numerical algorithms enable you to parallelize MATLAB® applications without CUDA or MPI programming. The toolbox lets you use parallel-enabled functions in MATLAB and other toolboxes. You can use the toolbox with Simulink® to run multiple simulations of a model in parallel. Programs and models can run in both interactive and batch modes.\n\n[Partial Differential Equation Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fpde.html) is a tool that provides functions for solving structural mechanics, heat transfer, and general partial differential equations (PDEs) using finite element analysis.\n\n[ROS Toolbox](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fros.html) is a tool that provides an interface connecting MATLAB® and Simulink® with the Robot Operating System (ROS and ROS 2), enabling you to create a network of ROS nodes. The toolbox includes MATLAB functions and Simulink blocks to import, analyze, and play back ROS data recorded in rosbag files. You can also connect to a live ROS network to access ROS messages.\n\n[Robotics Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Frobotics.html) provides a toolbox that brings robotics specific functionality(designing, simulating, and testing manipulators, mobile robots, and humanoid robots) to MATLAB, exploiting the native capabilities of MATLAB (linear algebra, portability, graphics). The toolbox also supports mobile robots with functions for robot motion models (bicycle), path planning algorithms (bug, distance transform, D*, PRM), kinodynamic planning (lattice, RRT), localization (EKF, particle filter), map building (EKF) and simultaneous localization and mapping (EKF), and a Simulink model a of non-holonomic vehicle. The Toolbox also including a detailed Simulink model for a quadrotor flying robot.\n\n[Deep Learning Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdeep-learning.html) is a tool that provides a framework for designing and implementing deep neural networks with algorithms, pretrained models, and apps. You can use convolutional neural networks (ConvNets, CNNs) and long short-term memory (LSTM) networks to perform classification and regression on image, time-series, and text data. You can build network architectures such as generative adversarial networks (GANs) and Siamese networks using automatic differentiation, custom training loops, and shared weights. With the Deep Network Designer app, you can design, analyze, and train networks graphically. It can exchange models with TensorFlow™ and PyTorch through the ONNX format and import models from TensorFlow-Keras and Caffe. The toolbox supports transfer learning with DarkNet-53, ResNet-50, NASNet, SqueezeNet and many other pretrained models.\n\n[Reinforcement Learning Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Freinforcement-learning.html) is a tool that provides an app, functions, and a Simulink® block for training policies using reinforcement learning algorithms, including DQN, PPO, SAC, and DDPG. You can use these policies to implement controllers and decision-making algorithms for complex applications such as resource allocation, robotics, and autonomous systems.\n\n[Deep Learning HDL Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdeep-learning-hdl.html) is a tool that provides functions and tools to prototype and implement deep learning networks on FPGAs and SoCs. It provides pre-built bitstreams for running a variety of deep learning networks on supported Xilinx® and Intel® FPGA and SoC devices. Profiling and estimation tools let you customize a deep learning network by exploring design, performance, and resource utilization tradeoffs.\n\n[Model Predictive Control Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmodel-predictive-control.html) is a tool that provides functions, an app, and Simulink® blocks for designing and simulating controllers using linear and nonlinear model predictive control (MPC). The toolbox lets you specify plant and disturbance models, horizons, constraints, and weights. By running closed-loop simulations, you can evaluate controller performance.\n\n[Vision HDL Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fvision-hdl.html) is a tool that provides pixel-streaming algorithms for the design and implementation of vision systems on FPGAs and ASICs. It provides a design framework that supports a diverse set of interface types, frame sizes, and frame rates. The image processing, video, and computer vision algorithms in the toolbox use an architecture appropriate for HDL implementations.\n\n[Data Acquisition Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdata-acquisition.html) is a tool that provides apps and functions for configuring data acquisition hardware, reading data into MATLAB® and Simulink®, and writing data to DAQ analog and digital output channels. The toolbox supports a variety of DAQ hardware, including USB, PCI, PCI Express®, PXI®, and PXI Express® devices, from National Instruments® and other vendors.\n\n[Microsoft AirSim](https:\u002F\u002Fmicrosoft.github.io\u002FAirSim\u002Flidar.html) is a simulator for drones, cars and more, built on Unreal Engine (with an experimental Unity release). AirSim is open-source, cross platform, and supports [software-in-the-loop simulation](https:\u002F\u002Fwww.mathworks.com\u002Fhelp\u002F\u002F\u002Fecoder\u002Fsoftware-in-the-loop-sil-simulation.html) with popular flight controllers such as PX4 & ArduPilot and [hardware-in-loop](https:\u002F\u002Fwww.ni.com\u002Fen-us\u002Finnovations\u002Fwhite-papers\u002F17\u002Fwhat-is-hardware-in-the-loop-.html) with PX4 for physically and visually realistic simulations. It is developed as an Unreal plugin that can simply be dropped into any Unreal environment. AirSim is being developed  as a platform for AI research to experiment with deep learning, computer vision and reinforcement learning algorithms for autonomous vehicles.\n\n# NLP Development\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_c20b531e32f3.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## NLP Learning Resources\n\n[Natural Language Processing (NLP)](https:\u002F\u002Fwww.ibm.com\u002Fcloud\u002Flearn\u002Fnatural-language-processing) is a branch of artificial intelligence (AI) focused on giving computers the ability to understand text and spoken words in much the same way human beings can. NLP combines computational linguistics rule-based modeling of human language with statistical, machine learning, and deep learning models.\n\n[Natural Language Processing With Python's NLTK Package](https:\u002F\u002Frealpython.com\u002Fnltk-nlp-python\u002F)\n\n[Cognitive Services—APIs for AI Developers | Microsoft Azure](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fcognitive-services\u002F)\n\n[Artificial Intelligence Services - Amazon Web Services (AWS)](https:\u002F\u002Faws.amazon.com\u002Fmachine-learning\u002Fai-services\u002F)\n\n[Google Cloud Natural Language API](https:\u002F\u002Fcloud.google.com\u002Fnatural-language\u002Fdocs\u002Freference\u002Frest)\n\n[Top Natural Language Processing Courses Online | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fnatural-language-processing\u002F)\n\n[Introduction to Natural Language Processing (NLP) | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fnatural-language-processing\u002F)\n\n[Top Natural Language Processing Courses | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?=&query=natural%20language%20processing)\n\n[Natural Language Processing | Coursera](https:\u002F\u002Fwww.coursera.org\u002Flearn\u002Flanguage-processing)\n\n[Natural Language Processing in TensorFlow | Coursera](https:\u002F\u002Fwww.coursera.org\u002Flearn\u002Fnatural-language-processing-tensorflow)\n\n[Learn Natural Language Processing with Online Courses and Lessons | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fnatural-language-processing)\n\n[Build a Natural Language Processing Solution with Microsoft Azure | Pluralsight](https:\u002F\u002Fwww.pluralsight.com\u002Fcourses\u002Fbuild-natural-language-processing-solution-microsoft-azure)\n\n[Natural Language Processing (NLP) Training Courses | NobleProg](https:\u002F\u002Fwww.nobleprog.com\u002Fnlp-training)\n\n[Natural Language Processing with Deep Learning Course | Standford Online](https:\u002F\u002Fonline.stanford.edu\u002Fcourses\u002Fcs224n-natural-language-processing-deep-learning)\n\n[Advanced Natural Language Processing - MIT OpenCourseWare](https:\u002F\u002Focw.mit.edu\u002Fcourses\u002Felectrical-engineering-and-computer-science\u002F6-864-advanced-natural-language-processing-fall-2005\u002F)\n\n[Certified Natural Language Processing Expert Certification | IABAC](https:\u002F\u002Fiabac.org\u002Fartificial-intelligence-certification\u002Fcertified-natural-language-processing-expert\u002F)\n\n[Natural Language Processing Course - Intel](https:\u002F\u002Fsoftware.intel.com\u002Fcontent\u002Fwww\u002Fus\u002Fen\u002Fdevelop\u002Ftraining\u002Fcourse-natural-language-processing.html)\n\n\n## NLP Tools, Libraries, and Frameworks\n\n[Natural Language Toolkit (NLTK)](https:\u002F\u002Fwww.nltk.org\u002F) is a leading platform for building Python programs to work with human language data. It provides easy-to-use interfaces to over [50 corpora and lexical resources](https:\u002F\u002Fnltk.org\u002Fnltk_data\u002F) such as WordNet, along with a suite of text processing libraries for classification, tokenization, stemming, tagging, parsing, and semantic reasoning, wrappers for industrial-strength NLP libraries.\n\n[spaCy](https:\u002F\u002Fspacy.io) is a library for advanced Natural Language Processing in Python and Cython. It's built on the very latest research, and was designed from day one to be used in real products. spaCy comes with pretrained pipelines and currently supports tokenization and training for 60+ languages. It also features neural network models for tagging, parsing, named entity recognition, text classification and more, multi-task learning with pretrained transformers like BERT.\n\n[CoreNLP](https:\u002F\u002Fstanfordnlp.github.io\u002FCoreNLP\u002F) is a set of natural language analysis tools written in Java. CoreNLP enables users to derive linguistic annotations for text, including token and sentence boundaries, parts of speech, named entities, numeric and time values, dependency and constituency parses, coreference, sentiment, quote attributions, and relations.\n\n[NLPnet](https:\u002F\u002Fgithub.com\u002Ferickrf\u002Fnlpnet) is a Python library for Natural Language Processing tasks based on neural networks. It performs part-of-speech tagging, semantic role labeling and dependency parsing.\n\n[Flair](https:\u002F\u002Fgithub.com\u002FflairNLP\u002Fflair) is a simple framework for state-of-the-art Natural Language Processing (NLP) models to your text, such as named entity recognition (NER), part-of-speech tagging (PoS), special support for biomedical data, sense disambiguation and classification, with support for a rapidly growing number of languages.\n\n[Catalyst](https:\u002F\u002Fgithub.com\u002Fcuriosity-ai\u002Fcatalyst) is a C# Natural Language Processing library built for speed. Inspired by [spaCy's design](https:\u002F\u002Fspacy.io\u002F), it brings pre-trained models, out-of-the box support for training word and document embeddings, and flexible entity recognition models.\n\n[Apache OpenNLP](https:\u002F\u002Fopennlp.apache.org\u002F) is an open-source library for a machine learning based toolkit used in the processing of natural language text. It features an API for use cases like [Named Entity Recognition](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FNamed-entity_recognition), [Sentence Detection](), [POS(Part-Of-Speech) tagging](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FPart-of-speech_tagging), [Tokenization](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FTokenization_(data_security)) [Feature extraction](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FFeature_extraction), [Chunking](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FChunking_(psychology)), [Parsing](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FParsing), and [Coreference resolution](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FCoreference).\n\n[Microsoft Cognitive Toolkit (CNTK)](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fcognitive-toolkit\u002F) is an open-source toolkit for commercial-grade distributed deep learning. It describes neural networks as a series of computational steps via a directed graph. CNTK allows the user to easily realize and combine popular model types such as feed-forward DNNs, convolutional neural networks (CNNs) and recurrent neural networks (RNNs\u002FLSTMs). CNTK implements stochastic gradient descent (SGD, error backpropagation) learning with automatic differentiation and parallelization across multiple GPUs and servers.\n\n[NVIDIA cuDNN](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcudnn) is a GPU-accelerated library of primitives for [deep neural networks](https:\u002F\u002Fdeveloper.nvidia.com\u002Fdeep-learning). cuDNN provides highly tuned implementations for standard routines such as forward and backward convolution, pooling, normalization, and activation layers. cuDNN accelerates widely used deep learning frameworks, including [Caffe2](https:\u002F\u002Fcaffe2.ai\u002F), [Chainer](https:\u002F\u002Fchainer.org\u002F), [Keras](https:\u002F\u002Fkeras.io\u002F), [MATLAB](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fdeep-learning.html), [MxNet](https:\u002F\u002Fmxnet.incubator.apache.org\u002F), [PyTorch](https:\u002F\u002Fpytorch.org\u002F), and [TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002F).\n\n[TensorFlow](https:\u002F\u002Fwww.tensorflow.org) is an end-to-end open source platform for machine learning. It has a comprehensive, flexible ecosystem of tools, libraries and community resources that lets researchers push the state-of-the-art in ML and developers easily build and deploy ML powered applications.\n\n[Tensorflow_macOS](https:\u002F\u002Fgithub.com\u002Fapple\u002Ftensorflow_macos) is a Mac-optimized version of TensorFlow and TensorFlow Addons for macOS 11.0+ accelerated using Apple's ML Compute framework.\n\n[Keras](https:\u002F\u002Fkeras.io) is a high-level neural networks API, written in Python and capable of running on top of TensorFlow, CNTK, or Theano.It was developed with a focus on enabling fast experimentation. It is capable of running on top of TensorFlow, Microsoft Cognitive Toolkit, R, Theano, or PlaidML.\n\n[PyTorch](https:\u002F\u002Fpytorch.org) is a library for deep learning on irregular input data such as graphs, point clouds, and manifolds. Primarily developed by Facebook's AI Research lab.\n\n[Eclipse Deeplearning4J (DL4J)](https:\u002F\u002Fdeeplearning4j.konduit.ai\u002F) is a set of projects intended to support all the needs of a JVM-based(Scala, Kotlin, Clojure, and Groovy) deep learning application. This means starting with the raw data, loading and preprocessing it from wherever and whatever format it is in to building and tuning a wide variety of simple and complex deep learning networks.\n\n[Chainer](https:\u002F\u002Fchainer.org\u002F) is a Python-based deep learning framework aiming at flexibility. It provides automatic differentiation APIs based on the define-by-run approach (dynamic computational graphs) as well as object-oriented high-level APIs to build and train neural networks. It also supports CUDA\u002FcuDNN using [CuPy](https:\u002F\u002Fgithub.com\u002Fcupy\u002Fcupy) for high performance training and inference.\n\n[Anaconda](https:\u002F\u002Fwww.anaconda.com\u002F) is a very popular Data Science platform for machine learning and deep learning that enables users to develop models, train them, and deploy them.\n\n[PlaidML](https:\u002F\u002Fgithub.com\u002Fplaidml\u002Fplaidml) is an advanced and portable tensor compiler for enabling deep learning on laptops, embedded devices, or other devices where the available computing hardware is not well supported or the available software stack contains unpalatable license restrictions.\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) is a Python module for machine learning built on top of SciPy, NumPy, and matplotlib, making it easier to apply robust and simple implementations of many popular machine learning algorithms.\n\n[Caffe](https:\u002F\u002Fgithub.com\u002FBVLC\u002Fcaffe) is a deep learning framework made with expression, speed, and modularity in mind. It is developed by Berkeley AI Research (BAIR)\u002FThe Berkeley Vision and Learning Center (BVLC) and community contributors.\n\n[Theano](https:\u002F\u002Fgithub.com\u002FTheano\u002FTheano) is a Python library that allows you to define, optimize, and evaluate mathematical expressions involving multi-dimensional arrays efficiently including tight integration with NumPy.\n\n[Apache Spark](https:\u002F\u002Fspark.apache.org\u002F) is a unified analytics engine for large-scale data processing. It provides high-level APIs in Scala, Java, Python, and R, and an optimized engine that supports general computation graphs for data analysis. It also supports a rich set of higher-level tools including Spark SQL for SQL and DataFrames, MLlib for machine learning, GraphX for graph processing, and Structured Streaming for stream processing.\n\n[Apache Spark Connector for SQL Server and Azure SQL](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fsql-spark-connector) is a high-performance connector that enables you to use transactional data in big data analytics and persists results for ad-hoc queries or reporting. The connector allows you to use any SQL database, on-premises or in the cloud, as an input data source or output data sink for Spark jobs.\n\n[Apache PredictionIO](https:\u002F\u002Fpredictionio.apache.org\u002F) is an open source machine learning framework for developers, data scientists, and end users. It supports event collection, deployment of algorithms, evaluation, querying predictive results via REST APIs. It is based on scalable open source services like Hadoop, HBase (and other DBs), Elasticsearch, Spark and implements what is called a Lambda Architecture.\n\n[Apache Airflow](https:\u002F\u002Fairflow.apache.org) is an open-source workflow management platform created by the community to programmatically author, schedule and monitor workflows. Airflow has a modular architecture and uses a message queue to orchestrate an arbitrary number of workers. Airflow is ready to scale to infinity.\n\n[Open Neural Network Exchange(ONNX)](https:\u002F\u002Fgithub.com\u002Fonnx) is an open ecosystem that empowers AI developers to choose the right tools as their project evolves. ONNX provides an open source format for AI models, both deep learning and traditional ML. It defines an extensible computation graph model, as well as definitions of built-in operators and standard data types.\n\n[BigDL](https:\u002F\u002Fbigdl-project.github.io\u002F) is a distributed deep learning library for Apache Spark. With BigDL, users can write their deep learning applications as standard Spark programs, which can directly run on top of existing Spark or Hadoop clusters.\n\n[Numba](https:\u002F\u002Fgithub.com\u002Fnumba\u002Fnumba) is an open source, NumPy-aware optimizing compiler for Python sponsored by Anaconda, Inc. It uses the LLVM compiler project to generate machine code from Python syntax. Numba can compile a large subset of numerically-focused Python, including many NumPy functions. Additionally, Numba has support for automatic parallelization of loops, generation of GPU-accelerated code, and creation of ufuncs and C callbacks.\n\n# Bioinformatics\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_cd7bb1386d0d.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## Bioinformatics Learning Resources\n\n[Bioinformatics](https:\u002F\u002Fwww.genome.gov\u002Fgenetics-glossary\u002FBioinformatics) is a field of computational science that has to do with the analysis of sequences of biological molecules. This usually refers to genes, DNA, RNA, or protein, and is particularly useful in comparing genes and other sequences in proteins and other sequences within an organism or between organisms, looking at evolutionary relationships between organisms, and using the patterns that exist across DNA and protein sequences to figure out what their function is.\n\n[European Bioinformatics Institute](https:\u002F\u002Fwww.ebi.ac.uk\u002F)\n\n[National Center for Biotechnology Information](https:\u002F\u002Fwww.ncbi.nlm.nih.gov)\n\n[Online Courses in Bioinformatics |ISCB - International Society for Computational Biology](https:\u002F\u002Fwww.iscb.org\u002Fcms_addon\u002Fonline_courses\u002Findex.php)\n\n[Bioinformatics | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fspecializations\u002Fbioinformatics)\n\n[Top Bioinformatics Courses | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002FBioinformatics\u002F)\n\n[Biometrics Courses | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fbiometrics\u002F)\n\n[Learn Bioinformatics with Online Courses and Lessons | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fbioinformatics)\n\n[Bioinformatics Graduate Certificate | Harvard Extension School](https:\u002F\u002Fextension.harvard.edu\u002Facademics\u002Fprograms\u002Fbioinformatics-graduate-certificate\u002F)\n\n[Bioinformatics and Biostatistics | UC San Diego Extension](https:\u002F\u002Fextension.ucsd.edu\u002Fcourses-and-programs\u002Fbioinformatics-and-biostatistics)\n\n[Bioinformatics and Proteomics - Free Online Course Materials | MIT](https:\u002F\u002Focw.mit.edu\u002Fcourses\u002Felectrical-engineering-and-computer-science\u002F6-092-bioinformatics-and-proteomics-january-iap-2005\u002F)\n\n[Introduction to Biometrics course - Biometrics Institute](https:\u002F\u002Fwww.biometricsinstitute.org\u002Fevent\u002Fintroduction-to-biometrics-short-course\u002F)\n\n## Bioinformatics Tools, Libraries, and Frameworks\n\n[Bioconductor](https:\u002F\u002Fbioconductor.org\u002F) is an open source project that provides tools for the analysis and comprehension of high-throughput genomic data. Bioconductor uses the [R statistical programming language](https:\u002F\u002Fwww.r-project.org\u002Fabout.html), and is open source and open development. It has two releases each year, and an active user community. Bioconductor is also available as an [AMI (Amazon Machine Image)](https:\u002F\u002Fdocs.aws.amazon.com\u002FAWSEC2\u002Flatest\u002FUserGuide\u002FAMIs.html) and [Docker images](https:\u002F\u002Fdocs.docker.com\u002Fengine\u002Freference\u002Fcommandline\u002Fimages\u002F).\n\n[Bioconda](https:\u002F\u002Fbioconda.github.io) is a channel for the conda package manager specializing in bioinformatics software. It has a repository of packages containing over 7000 bioinformatics packages ready to use with conda install.\n\n[UniProt](https:\u002F\u002Fwww.uniprot.org\u002F) is a freely accessible database that provide users with a comprehensive, high-quality and freely accessible set of protein sequences annotated with functional information.\n\n[Bowtie 2](https:\u002F\u002Fbio.tools\u002Fbowtie2#!) is an ultrafast and memory-efficient tool for aligning sequencing reads to long reference sequences. It is particularly good at aligning reads of about 50 up to 100s or 1,000s of characters, and particularly good at aligning to relatively long (mammalian) genomes.\n\n[Biopython](https:\u002F\u002Fbiopython.org\u002F) is a set of freely available tools for biological computation written in Python by an international team of developers. It is a distributed collaborative effort to develop Python libraries and applications which address the needs of current and future work in bioinformatics.\n\n[BioRuby](https:\u002F\u002Fbioruby.open-bio.org\u002F) is a toolkit that has components for sequence analysis, pathway analysis, protein modelling and phylogenetic analysis; it supports many widely used data formats and provides easy access to databases, external programs and public web services, including BLAST, KEGG, GenBank, MEDLINE and GO.\n\n[BioJava](https:\u002F\u002Fbiojava.org\u002F) is a toolkit that provides an API to maintain local installations of the PDB, load and manipulate structures, perform standard analysis such as sequence and structure alignments and visualize them in 3D.\n\n[BioPHP](https:\u002F\u002Fbiophp.org\u002F) is an open source project that provides a collection of open source PHP code, with classes for DNA and protein sequence analysis, alignment, database parsing, and other bioinformatics tools.\n\n[Avogadro](https:\u002F\u002Favogadro.cc\u002F) is an advanced molecule editor and visualizer designed for cross-platform use in computational chemistry, molecular modeling, bioinformatics, materials science, and related areas. It offers flexible high quality rendering and a powerful plugin architecture.\n\n[Ascalaph Designer](https:\u002F\u002Fwww.biomolecular-modeling.com\u002FAscalaph\u002FAscalaph_Designer.html) is a program for molecular dynamic simulations. Under a single graphical environment are represented as their own implementation of molecular dynamics as well as the methods of classical and quantum mechanics of popular programs.\n\n[Anduril](https:\u002F\u002Fwww.anduril.org\u002Fsite\u002F) is a workflow platform for analyzing large data sets. Anduril provides facilities for analyzing high-thoughput data in biomedical research, and the platform is fully extensible by third parties. Ready-made tools support data visualization, DNA\u002FRNA\u002FChIP-sequencing, DNA\u002FRNA microarrays, cytometry and image analysis.\n\n[Galaxy](https:\u002F\u002Fmelbournebioinformatics.github.io\u002FMelBioInf_docs\u002Ftutorials\u002Fgalaxy_101\u002Fgalaxy_101\u002F) is an open source, web-based platform for accessible, reproducible, and transparent computational biomedical research. It allows users without programming experience to easily specify parameters and run individual tools as well as larger workflows. It also captures run information so that any user can repeat and understand a complete computational analysis.\n\n[PathVisio](https:\u002F\u002Fpathvisio.github.io\u002F) is a free open-source pathway analysis and drawing software which allows drawing, editing, and analyzing biological pathways. It is developed in Java and can be extended with plugins.\n\n[Orange](https:\u002F\u002Forangedatamining.com\u002F) is a powerful data mining and machine learning toolkit that performs data analysis and visualization.\n\n[Basic Local Alignment Search Tool](https:\u002F\u002Fblast.ncbi.nlm.nih.gov\u002FBlast.cgi) is a tool that finds regions of similarity between biological sequences. The program compares nucleotide or protein sequences to sequence databases and calculates the statistical significance.\n\n[OSIRIS](https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fosiris\u002F) is public-domain, free, and open source STR analysis software designed for clinical, forensic, and research use, and has been validated for use as an expert system for single-source samples.\n\n[NCBI BioSystems](https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fbiosystems\u002F) is a  Database that provides integrated access to biological systems and their component genes, proteins, and small molecules, as well as literature describing those biosystems and other related data throughout Entrez.\n\n# CUDA Development\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_19a3d218f324.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_34ac80e10409.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**CUDA Toolkit. Source: [NVIDIA Developer CUDA](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcuda-zone)**\n\n## CUDA Learning Resources\n\n[CUDA](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcuda-zone) is a parallel computing platform and programming model developed by NVIDIA for general computing on graphical processing units (GPUs). With CUDA, developers are able to dramatically speed up computing applications by harnessing the power of GPUs. In GPU-accelerated applications, the sequential part of the workload runs on the CPU, which is optimized for single-threaded. The compute intensive portion of the application runs on thousands of GPU cores in parallel. When using CUDA, developers can program in popular languages such as C, C++, Fortran, Python and MATLAB.\n\n[CUDA Toolkit Documentation](https:\u002F\u002Fdocs.nvidia.com\u002Fcuda\u002Findex.html)\n\n[CUDA Quick Start Guide](https:\u002F\u002Fdocs.nvidia.com\u002Fcuda\u002Fcuda-quick-start-guide\u002Findex.html)\n\n[CUDA on WSL](https:\u002F\u002Fdocs.nvidia.com\u002Fcuda\u002Fwsl-user-guide\u002Findex.html)\n\n[CUDA GPU support for TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002Finstall\u002Fgpu)\n\n[NVIDIA Deep Learning cuDNN Documentation](https:\u002F\u002Fdocs.nvidia.com\u002Fdeeplearning\u002Fcudnn\u002Fapi\u002Findex.html)\n\n[NVIDIA GPU Cloud Documentation](https:\u002F\u002Fdocs.nvidia.com\u002Fngc\u002Fngc-introduction\u002Findex.html)\n\n[NVIDIA NGC](https:\u002F\u002Fngc.nvidia.com\u002F) is a hub for GPU-optimized software for deep learning, machine learning, and high-performance computing (HPC) workloads.\n\n[NVIDIA NGC Containers](https:\u002F\u002Fwww.nvidia.com\u002Fen-us\u002Fgpu-cloud\u002Fcontainers\u002F) is a registry that provides researchers, data scientists, and developers with simple access to a comprehensive catalog of GPU-accelerated software for AI, machine learning and HPC. These containers take full advantage of NVIDIA GPUs on-premises and in the cloud.\n\n## CUDA Tools Libraries, and Frameworks\n\n[CUDA Toolkit](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcuda-downloads) is a collection of tools & libraries that provide a development environment for creating high performance GPU-accelerated applications. The CUDA Toolkit allows you can develop, optimize, and deploy your applications on GPU-accelerated embedded systems, desktop workstations, enterprise data centers, cloud-based platforms and HPC supercomputers. The toolkit includes GPU-accelerated libraries, debugging and optimization tools, a C\u002FC++ compiler, and a runtime library to build and deploy your application on major architectures including x86, Arm and POWER.\n\n[NVIDIA cuDNN](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcudnn) is a GPU-accelerated library of primitives for [deep neural networks](https:\u002F\u002Fdeveloper.nvidia.com\u002Fdeep-learning). cuDNN provides highly tuned implementations for standard routines such as forward and backward convolution, pooling, normalization, and activation layers. cuDNN accelerates widely used deep learning frameworks, including [Caffe2](https:\u002F\u002Fcaffe2.ai\u002F), [Chainer](https:\u002F\u002Fchainer.org\u002F), [Keras](https:\u002F\u002Fkeras.io\u002F), [MATLAB](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fdeep-learning.html), [MxNet](https:\u002F\u002Fmxnet.incubator.apache.org\u002F), [PyTorch](https:\u002F\u002Fpytorch.org\u002F), and [TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002F).\n\n[CUDA-X HPC](https:\u002F\u002Fwww.nvidia.com\u002Fen-us\u002Ftechnologies\u002Fcuda-x\u002F) is a collection of libraries, tools, compilers and APIs that help developers solve the world's most challenging problems. CUDA-X HPC includes highly tuned kernels essential for high-performance computing (HPC).\n\n[NVIDIA Container Toolkit](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fnvidia-docker) is a collection of tools & libraries that allows users to build and run GPU accelerated Docker containers. The toolkit includes a container runtime [library](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Flibnvidia-container) and utilities to automatically configure containers to leverage NVIDIA GPUs.\n\n[Minkowski Engine](https:\u002F\u002Fnvidia.github.io\u002FMinkowskiEngine) is an auto-differentiation library for sparse tensors. It supports all standard neural network layers such as convolution, pooling, unpooling, and broadcasting operations for sparse tensors.\n\n[CUTLASS](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fcutlass) is a collection of CUDA C++ template abstractions for implementing high-performance matrix-multiplication (GEMM) at all levels and scales within CUDA. It incorporates strategies for hierarchical decomposition and data movement similar to those used to implement cuBLAS.\n\n[CUB](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fcub) is a cooperative primitives for CUDA C++ kernel authors.\n\n[Tensorman](https:\u002F\u002Fgithub.com\u002Fpop-os\u002Ftensorman) is a utility for easy management of Tensorflow containers by developed by [System76]( https:\u002F\u002Fsystem76.com).Tensorman allows Tensorflow to operate in an isolated environment that is contained from the rest of the system. This virtual environment can operate independent of the base system, allowing you to use any version of Tensorflow on any version of a Linux distribution that supports the Docker runtime.\n\n[Numba](https:\u002F\u002Fgithub.com\u002Fnumba\u002Fnumba) is an open source, NumPy-aware optimizing compiler for Python sponsored by Anaconda, Inc. It uses the LLVM compiler project to generate machine code from Python syntax. Numba can compile a large subset of numerically-focused Python, including many NumPy functions. Additionally, Numba has support for automatic parallelization of loops, generation of GPU-accelerated code, and creation of ufuncs and C callbacks.\n\n[Chainer](https:\u002F\u002Fchainer.org\u002F) is a Python-based deep learning framework aiming at flexibility. It provides automatic differentiation APIs based on the define-by-run approach (dynamic computational graphs) as well as object-oriented high-level APIs to build and train neural networks. It also supports CUDA\u002FcuDNN using [CuPy](https:\u002F\u002Fgithub.com\u002Fcupy\u002Fcupy) for high performance training and inference.\n\n[CuPy](https:\u002F\u002Fcupy.dev\u002F) is an implementation of NumPy-compatible multi-dimensional array on CUDA. CuPy consists of the core multi-dimensional array class, cupy.ndarray, and many functions on it. It supports a subset of numpy.ndarray interface.\n\n[CatBoost](https:\u002F\u002Fcatboost.ai\u002F) is a fast, scalable, high performance [Gradient Boosting](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FGradient_boosting) on Decision Trees library, used for ranking, classification, regression and other machine learning tasks for Python, R, Java, C++. Supports computation on CPU and GPU.\n\n[cuDF](https:\u002F\u002Frapids.ai\u002F) is a GPU DataFrame library for loading, joining, aggregating, filtering, and otherwise manipulating data. cuDF provides a pandas-like API that will be familiar to data engineers & data scientists, so they can use it to easily accelerate their workflows without going into the details of CUDA programming.\n\n[cuML](https:\u002F\u002Fgithub.com\u002Frapidsai\u002Fcuml) is a suite of libraries that implement machine learning algorithms and mathematical primitives functions that share compatible APIs with other RAPIDS projects. cuML enables data scientists, researchers, and software engineers to run traditional tabular ML tasks on GPUs without going into the details of CUDA programming. In most cases, cuML's Python API matches the API from scikit-learn.\n\n[ArrayFire](https:\u002F\u002Farrayfire.com\u002F) is a general-purpose library that simplifies the process of developing software that targets parallel and massively-parallel architectures including CPUs, GPUs, and other hardware acceleration devices.\n\n[Thrust](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fthrust) is a C++ parallel programming library which resembles the C++ Standard Library. Thrust's high-level interface greatly enhances programmer productivity while enabling performance portability between GPUs and multicore CPUs.\n\n[AresDB](https:\u002F\u002Feng.uber.com\u002Faresdb\u002F) is a GPU-powered real-time analytics storage and query engine. It features low query latency, high data freshness and highly efficient in-memory and on disk storage management.\n\n[Arraymancer](https:\u002F\u002Fmratsim.github.io\u002FArraymancer\u002F) is a tensor (N-dimensional array) project in Nim. The main focus is providing a fast and ergonomic CPU, Cuda and OpenCL ndarray library on which to build a scientific computing ecosystem.\n\n[Kintinuous](https:\u002F\u002Fgithub.com\u002Fmp3guy\u002FKintinuous) is a real-time dense visual SLAM system capable of producing high quality globally consistent point and mesh reconstructions over hundreds of metres in real-time with only a low-cost commodity RGB-D sensor.\n\n[GraphVite](https:\u002F\u002Fgraphvite.io\u002F) is a general graph embedding engine, dedicated to high-speed and large-scale embedding learning in various applications.\n\n# MATLAB Development\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_5bde272346f3.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## MATLAB Learning Resources\n\n[MATLAB](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmatlab.html) is a programming language that does numerical computing such as expressing matrix and array mathematics directly.\n\n[MATLAB Documentation](https:\u002F\u002Fwww.mathworks.com\u002Fhelp\u002Fmatlab\u002F)\n\n[Getting Started with MATLAB ](https:\u002F\u002Fwww.mathworks.com\u002Fhelp\u002Fmatlab\u002Fgetting-started-with-matlab.html)\n\n[MATLAB and Simulink Training from MATLAB Academy](https:\u002F\u002Fmatlabacademy.mathworks.com)\n\n[MathWorks Certification Program](https:\u002F\u002Fwww.mathworks.com\u002Fservices\u002Ftraining\u002Fcertification.html)\n\n[MATLAB Online Courses from Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fmatlab\u002F)\n\n[MATLAB Online Courses from Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=matlab)\n\n[MATLAB Online Courses from edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fmatlab)\n\n[Building a MATLAB GUI](https:\u002F\u002Fwww.mathworks.com\u002Fdiscovery\u002Fmatlab-gui.html)\n\n[MATLAB Style Guidelines 2.0](https:\u002F\u002Fwww.mathworks.com\u002Fmatlabcentral\u002Ffileexchange\u002F46056-matlab-style-guidelines-2-0)\n\n[Setting Up Git Source Control with MATLAB & Simulink](https:\u002F\u002Fwww.mathworks.com\u002Fhelp\u002Fmatlab\u002Fmatlab_prog\u002Fset-up-git-source-control.html)\n\n[Pull, Push and Fetch Files with Git with MATLAB & Simulink](https:\u002F\u002Fwww.mathworks.com\u002Fhelp\u002Fmatlab\u002Fmatlab_prog\u002Fpush-and-fetch-with-git.html)\n\n[Create New Repository with MATLAB & Simulink](https:\u002F\u002Fwww.mathworks.com\u002Fhelp\u002Fmatlab\u002Fmatlab_prog\u002Fadd-folder-to-source-control.html)\n\n[PRMLT](http:\u002F\u002Fprml.github.io\u002F) is Matlab code for machine learning algorithms in the PRML book.\n\n## MATLAB Tools, Libraries, Frameworks\n\n**[MATLAB and Simulink Services & Applications List](https:\u002F\u002Fwww.mathworks.com\u002Fproducts.html)**\n\n[MATLAB in the Cloud](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fcloud.html) is a service that allows you to run in cloud environments from [MathWorks Cloud](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fcloud.html#browser) to [Public Clouds](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fcloud.html#public-cloud) including [AWS](https:\u002F\u002Faws.amazon.com\u002F) and [Azure](https:\u002F\u002Fazure.microsoft.com\u002F).\n\n[MATLAB Online™](https:\u002F\u002Fmatlab.mathworks.com) is a service that allows to users to uilitize MATLAB and Simulink through a web browser such as Google Chrome.\n\n[Simulink](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fsimulink.html) is a block diagram environment for Model-Based Design. It supports simulation, automatic code generation, and continuous testing of embedded systems.\n\n[Simulink Online™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fsimulink-online.html) is a service that provides access to Simulink through your web browser.\n\n[MATLAB Drive™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmatlab-drive.html) is a service that gives you the ability to store, access, and work with your files from anywhere.\n\n[MATLAB Parallel Server™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmatlab-parallel-server.html) is a tool that lets you scale MATLAB® programs and Simulink® simulations to clusters and clouds. You can prototype your programs and simulations on the desktop and then run them on clusters and clouds without recoding. MATLAB Parallel Server supports batch jobs, interactive parallel computations, and distributed computations with large matrices.\n\n[MATLAB Schemer](https:\u002F\u002Fgithub.com\u002Fscottclowe\u002Fmatlab-schemer) is a MATLAB package makes it easy to change the color scheme (theme) of the MATLAB display and GUI.\n\n[LRSLibrary](https:\u002F\u002Fgithub.com\u002Fandrewssobral\u002Flrslibrary) is a Low-Rank and Sparse Tools for Background Modeling and Subtraction in Videos. The library was designed for moving object detection in videos, but it can be also used for other computer vision and machine learning problems.\n\n[Image Processing Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fimage.html) is a tool that provides a comprehensive set of reference-standard algorithms and workflow apps for image processing, analysis, visualization, and algorithm development. You can perform image segmentation, image enhancement, noise reduction, geometric transformations, image registration, and 3D image processing.\n\n[Computer Vision Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fcomputer-vision.html) is a tool that provides algorithms, functions, and apps for designing and testing computer vision, 3D vision, and video processing systems. You can perform object detection and tracking, as well as feature detection, extraction, and matching. You can automate calibration workflows for single, stereo, and fisheye cameras. For 3D vision, the toolbox supports visual and point cloud SLAM, stereo vision, structure from motion, and point cloud processing.\n\n[Statistics and Machine Learning Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fstatistics.html) is a tool that provides functions and apps to describe, analyze, and model data. You can use descriptive statistics, visualizations, and clustering for exploratory data analysis; fit probability distributions to data; generate random numbers for Monte Carlo simulations, and perform hypothesis tests. Regression and classification algorithms let you draw inferences from data and build predictive models either interactively, using the Classification and Regression Learner apps, or programmatically, using AutoML.\n\n[Lidar Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Flidar.html) is a tool that provides algorithms, functions, and apps for designing, analyzing, and testing lidar processing systems. You can perform object detection and tracking, semantic segmentation, shape fitting, lidar registration, and obstacle detection. Lidar Toolbox supports lidar-camera cross calibration for workflows that combine computer vision and lidar processing.\n\n[Mapping Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmapping.html) is a tool that provides algorithms and functions for transforming geographic data and creating map displays. You can visualize your data in a geographic context, build map displays from more than 60 map projections, and transform data from a variety of sources into a consistent geographic coordinate system.\n\n[UAV Toolbox](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fuav.html) is an application that provides tools and reference applications for designing, simulating, testing, and deploying unmanned aerial vehicle (UAV) and drone applications. You can design autonomous flight algorithms, UAV missions, and flight controllers. The Flight Log Analyzer app lets you interactively analyze 3D flight paths, telemetry information, and sensor readings from common flight log formats.\n\n[Parallel Computing Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmatlab-parallel-server.html) is a tool that lets you solve computationally and data-intensive problems using multicore processors, GPUs, and computer clusters. High-level constructs such as parallel for-loops, special array types, and parallelized numerical algorithms enable you to parallelize MATLAB® applications without CUDA or MPI programming. The toolbox lets you use parallel-enabled functions in MATLAB and other toolboxes. You can use the toolbox with Simulink® to run multiple simulations of a model in parallel. Programs and models can run in both interactive and batch modes.\n\n[Partial Differential Equation Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fpde.html) is a tool that provides functions for solving structural mechanics, heat transfer, and general partial differential equations (PDEs) using finite element analysis.\n\n[ROS Toolbox](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fros.html) is a tool that provides an interface connecting MATLAB® and Simulink® with the Robot Operating System (ROS and ROS 2), enabling you to create a network of ROS nodes. The toolbox includes MATLAB functions and Simulink blocks to import, analyze, and play back ROS data recorded in rosbag files. You can also connect to a live ROS network to access ROS messages.\n\n[Robotics Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Frobotics.html) provides a toolbox that brings robotics specific functionality(designing, simulating, and testing manipulators, mobile robots, and humanoid robots) to MATLAB, exploiting the native capabilities of MATLAB (linear algebra, portability, graphics). The toolbox also supports mobile robots with functions for robot motion models (bicycle), path planning algorithms (bug, distance transform, D*, PRM), kinodynamic planning (lattice, RRT), localization (EKF, particle filter), map building (EKF) and simultaneous localization and mapping (EKF), and a Simulink model a of non-holonomic vehicle. The Toolbox also including a detailed Simulink model for a quadrotor flying robot.\n\n[Deep Learning Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdeep-learning.html) is a tool that provides a framework for designing and implementing deep neural networks with algorithms, pretrained models, and apps. You can use convolutional neural networks (ConvNets, CNNs) and long short-term memory (LSTM) networks to perform classification and regression on image, time-series, and text data. You can build network architectures such as generative adversarial networks (GANs) and Siamese networks using automatic differentiation, custom training loops, and shared weights. With the Deep Network Designer app, you can design, analyze, and train networks graphically. It can exchange models with TensorFlow™ and PyTorch through the ONNX format and import models from TensorFlow-Keras and Caffe. The toolbox supports transfer learning with DarkNet-53, ResNet-50, NASNet, SqueezeNet and many other pretrained models.\n\n[Reinforcement Learning Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Freinforcement-learning.html) is a tool that provides an app, functions, and a Simulink® block for training policies using reinforcement learning algorithms, including DQN, PPO, SAC, and DDPG. You can use these policies to implement controllers and decision-making algorithms for complex applications such as resource allocation, robotics, and autonomous systems.\n\n[Deep Learning HDL Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdeep-learning-hdl.html) is a tool that provides functions and tools to prototype and implement deep learning networks on FPGAs and SoCs. It provides pre-built bitstreams for running a variety of deep learning networks on supported Xilinx® and Intel® FPGA and SoC devices. Profiling and estimation tools let you customize a deep learning network by exploring design, performance, and resource utilization tradeoffs.\n\n[Model Predictive Control Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmodel-predictive-control.html) is a tool that provides functions, an app, and Simulink® blocks for designing and simulating controllers using linear and nonlinear model predictive control (MPC). The toolbox lets you specify plant and disturbance models, horizons, constraints, and weights. By running closed-loop simulations, you can evaluate controller performance.\n\n[Vision HDL Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fvision-hdl.html) is a tool that provides pixel-streaming algorithms for the design and implementation of vision systems on FPGAs and ASICs. It provides a design framework that supports a diverse set of interface types, frame sizes, and frame rates. The image processing, video, and computer vision algorithms in the toolbox use an architecture appropriate for HDL implementations.\n\n[SoC Blockset™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fsoc.html) is a tool that provides Simulink® blocks and visualization tools for modeling, simulating, and analyzing hardware and software architectures for ASICs, FPGAs, and systems on a chip (SoC). You can build your system architecture using memory models, bus models, and I\u002FO models, and simulate the architecture together with the algorithms.\n\n[Wireless HDL Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fwireless-hdl.html) is a tool that provides pre-verified, hardware-ready Simulink® blocks and subsystems for developing 5G, LTE, and custom OFDM-based wireless communication applications. It includes reference applications, IP blocks, and gateways between frame and sample-based processing.\n\n[ThingSpeak™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fthingspeak.html) is an IoT analytics service that allows you to aggregate, visualize, and analyze live data streams in the cloud. ThingSpeak provides instant visualizations of data posted by your devices to ThingSpeak. With the ability to execute MATLAB® code in ThingSpeak, you can perform online analysis and process data as it comes in. ThingSpeak is often used for prototyping and proof-of-concept IoT systems that require analytics.\n\n[SEA-MAT](https:\u002F\u002Fsea-mat.github.io\u002Fsea-mat\u002F) is a collaborative effort to organize and distribute Matlab tools for the Oceanographic Community.\n\n[Gramm](https:\u002F\u002Fgithub.com\u002Fpiermorel\u002Fgramm) is a complete data visualization toolbox for Matlab. It provides an easy to use and high-level interface to produce publication-quality plots of complex data with varied statistical visualizations. Gramm is inspired by R's ggplot2 library.\n\n[hctsa](https:\u002F\u002Fhctsa-users.gitbook.io\u002Fhctsa-manual) is a software package for running highly comparative time-series analysis using Matlab.\n\n[Plotly](https:\u002F\u002Fplot.ly\u002Fmatlab\u002F) is a Graphing Library for MATLAB.\n\n[YALMIP](https:\u002F\u002Fyalmip.github.io\u002F) is a MATLAB toolbox for optimization modeling.\n\n[GNU Octave](https:\u002F\u002Fwww.gnu.org\u002Fsoftware\u002Foctave\u002F) is a high-level interpreted language, primarily intended for numerical computations. It provides capabilities for the numerical solution of linear and nonlinear problems, and for performing other numerical experiments. It also provides extensive graphics capabilities for data visualization and manipulation.\n\n# C\u002FC++ Development\n\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_24148fc60257.png\">\n  \u003Cbr \u002F>\n  \n\u003C\u002Fp>\n\n## C\u002FC++ Learning Resources\n\n[C++](https:\u002F\u002Fwww.cplusplus.com\u002Fdoc\u002Ftutorial\u002F) is a cross-platform language that can be used to build high-performance applications developed by Bjarne Stroustrup, as an extension to the C language.\n\n[C](https:\u002F\u002Fwww.iso.org\u002Fstandard\u002F74528.html) is a general-purpose, high-level language that was originally developed by Dennis M. Ritchie to develop the UNIX operating system at Bell Labs. It supports structured programming, lexical variable scope, and recursion, with a static type system. C also provides constructs that map efficiently to typical machine instructions, which makes it one was of the most widely used programming languages today.\n\n[Embedded C](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FEmbedded_C) is a set of language extensions for the C programming language by the [C Standards Committee](https:\u002F\u002Fisocpp.org\u002Fstd\u002Fthe-committee) to address issues that exist between C extensions for different [embedded systems](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FEmbedded_system). The extensions hep enhance microprocessor features such as fixed-point arithmetic, multiple distinct memory banks, and basic I\u002FO operations. This makes Embedded C the most popular embedded software language in the world.\n\n[C & C++ Developer Tools from JetBrains](https:\u002F\u002Fwww.jetbrains.com\u002Fcpp\u002F)\n\n[Open source C++ libraries on cppreference.com](https:\u002F\u002Fen.cppreference.com\u002Fw\u002Fcpp\u002Flinks\u002Flibs)\n\n[C++ Graphics libraries](https:\u002F\u002Fcpp.libhunt.com\u002Flibs\u002Fgraphics)\n\n[C++ Libraries in MATLAB](https:\u002F\u002Fwww.mathworks.com\u002Fhelp\u002Fmatlab\u002Fcall-cpp-library-functions.html)\n\n[C++ Tools and Libraries Articles](https:\u002F\u002Fwww.cplusplus.com\u002Farticles\u002Ftools\u002F)\n\n[Google C++ Style Guide](https:\u002F\u002Fgoogle.github.io\u002Fstyleguide\u002Fcppguide.html)\n\n[Introduction C++ Education course on Google Developers](https:\u002F\u002Fdevelopers.google.com\u002Fedu\u002Fc++\u002F)\n\n[C++ style guide for Fuchsia](https:\u002F\u002Ffuchsia.dev\u002Ffuchsia-src\u002Fdevelopment\u002Flanguages\u002Fc-cpp\u002Fcpp-style)\n\n[C and C++ Coding Style Guide by OpenTitan](https:\u002F\u002Fdocs.opentitan.org\u002Fdoc\u002Frm\u002Fc_cpp_coding_style\u002F)\n\n[Chromium C++ Style Guide](https:\u002F\u002Fchromium.googlesource.com\u002Fchromium\u002Fsrc\u002F+\u002Fmaster\u002Fstyleguide\u002Fc++\u002Fc++.md)\n\n[C++ Core Guidelines](https:\u002F\u002Fgithub.com\u002Fisocpp\u002FCppCoreGuidelines\u002Fblob\u002Fmaster\u002FCppCoreGuidelines.md)\n\n[C++ Style Guide for ROS](http:\u002F\u002Fwiki.ros.org\u002FCppStyleGuide)\n\n[Learn C++](https:\u002F\u002Fwww.learncpp.com\u002F)\n\n[Learn C : An Interactive C Tutorial](https:\u002F\u002Fwww.learn-c.org\u002F)\n\n[C++ Institute](https:\u002F\u002Fcppinstitute.org\u002Ffree-c-and-c-courses)\n\n[C++ Online Training Courses on LinkedIn Learning](https:\u002F\u002Fwww.linkedin.com\u002Flearning\u002Ftopics\u002Fc-plus-plus)\n\n[C++ Tutorials on W3Schools](https:\u002F\u002Fwww.w3schools.com\u002Fcpp\u002Fdefault.asp)\n\n[Learn C Programming Online Courses on edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fc-programming)\n\n[Learn C++ with Online Courses on edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fc-plus-plus)\n\n[Learn C++ on Codecademy](https:\u002F\u002Fwww.codecademy.com\u002Flearn\u002Flearn-c-plus-plus)\n\n[Coding for Everyone: C and C++ course on Coursera](https:\u002F\u002Fwww.coursera.org\u002Fspecializations\u002Fcoding-for-everyone)\n\n[C++ For C Programmers on Coursera](https:\u002F\u002Fwww.coursera.org\u002Flearn\u002Fc-plus-plus-a)\n\n[Top C Courses on Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=c%20programming)\n\n[C++ Online Courses on Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fc-plus-plus\u002F)\n\n[Top C Courses on Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fc-programming\u002F)\n\n[Basics of Embedded C Programming for Beginners on Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fembedded-c-programming-for-embedded-systems\u002F)\n\n[C++ For Programmers Course on Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fc-for-programmers--ud210)\n\n[C++ Fundamentals Course on Pluralsight](https:\u002F\u002Fwww.pluralsight.com\u002Fcourses\u002Flearn-program-cplusplus)\n\n[Introduction to C++ on MIT Free Online Course Materials](https:\u002F\u002Focw.mit.edu\u002Fcourses\u002Felectrical-engineering-and-computer-science\u002F6-096-introduction-to-c-january-iap-2011\u002F)\n\n[Introduction to C++ for Programmers | Harvard ](https:\u002F\u002Fonline-learning.harvard.edu\u002Fcourse\u002Fintroduction-c-programmers)\n\n[Online C Courses | Harvard University](https:\u002F\u002Fonline-learning.harvard.edu\u002Fsubject\u002Fc)\n\n\n## C\u002FC++ Tools\n\n[AWS SDK for C++](https:\u002F\u002Faws.amazon.com\u002Fsdk-for-cpp\u002F)\n\n[Azure SDK for C++](https:\u002F\u002Fgithub.com\u002FAzure\u002Fazure-sdk-for-cpp)\n\n[Azure SDK for C](https:\u002F\u002Fgithub.com\u002FAzure\u002Fazure-sdk-for-c)\n\n[C++ Client Libraries for Google Cloud Services](https:\u002F\u002Fgithub.com\u002Fgoogleapis\u002Fgoogle-cloud-cpp)\n\n[Visual Studio](https:\u002F\u002Fvisualstudio.microsoft.com\u002F) is an integrated development environment (IDE) from Microsoft; which is a feature-rich application that can be used for many aspects of software development. Visual Studio makes it easy to edit, debug, build, and publish your app. By using Microsoft software development platforms such as Windows API, Windows Forms, Windows Presentation Foundation, and Windows Store.\n\n[Visual Studio Code](https:\u002F\u002Fcode.visualstudio.com\u002F) is a code editor redefined and optimized for building and debugging modern web and cloud applications.\n\n[Vcpkg](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fvcpkg) is a C++ Library Manager for Windows, Linux, and MacOS.\n\n[ReSharper C++](https:\u002F\u002Fwww.jetbrains.com\u002Fresharper-cpp\u002Ffeatures\u002F) is a Visual Studio Extension for C++ developers developed by JetBrains.\n\n[AppCode](https:\u002F\u002Fwww.jetbrains.com\u002Fobjc\u002F) is constantly monitoring the quality of your code. It warns you of errors and smells and suggests quick-fixes to resolve them automatically. AppCode provides lots of code inspections for Objective-C, Swift, C\u002FC++, and a number of code inspections for other supported languages. All code inspections are run on the fly.\n\n[CLion](https:\u002F\u002Fwww.jetbrains.com\u002Fclion\u002Ffeatures\u002F) is a cross-platform IDE for C and C++ developers developed by JetBrains.\n\n[Code::Blocks](https:\u002F\u002Fwww.codeblocks.org\u002F) is a free C\u002FC++ and Fortran IDE built to meet the most demanding needs of its users. It is designed to be very extensible and fully configurable. Built around a plugin framework, Code::Blocks can be extended with plugins.\n\n[CppSharp](https:\u002F\u002Fgithub.com\u002Fmono\u002FCppSharp) is a tool and set of libraries which facilitates the usage of native C\u002FC++ code with the .NET ecosystem. It consumes C\u002FC++ header and library files and generates the necessary glue code to surface the native API as a managed API. Such an API can be used to consume an existing native library in your managed code or add managed scripting support to a native codebase.\n\n[Conan](https:\u002F\u002Fconan.io\u002F) is an Open Source Package Manager for C++ development and dependency management into the 21st century and on par with the other development ecosystems. \n\n[High Performance Computing (HPC) SDK](https:\u002F\u002Fdeveloper.nvidia.com\u002Fhpc) is a comprehensive toolbox for GPU accelerating HPC modeling and simulation applications. It includes the C, C++, and Fortran compilers, libraries, and analysis tools necessary for developing HPC applications on the NVIDIA platform.\n\n[Thrust](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fthrust) is a C++ parallel programming library which resembles the C++ Standard Library. Thrust's high-level interface greatly enhances programmer productivity while enabling performance portability between GPUs and multicore CPUs. Interoperability with established technologies such as CUDA, TBB, and OpenMP integrates with existing software.\n\n[Boost](https:\u002F\u002Fwww.boost.org\u002F) is an educational opportunity focused on cutting-edge C++. Boost has been a participant in the annual Google Summer of Code since 2007, in which students develop their skills by working on Boost Library development.\n\n[Automake](https:\u002F\u002Fwww.gnu.org\u002Fsoftware\u002Fautomake\u002F) is a tool for automatically generating Makefile.in files compliant with the GNU Coding Standards. Automake requires the use of GNU Autoconf.\n\n[Cmake](https:\u002F\u002Fcmake.org\u002F) is an open-source, cross-platform family of tools designed to build, test and package software. CMake is used to control the software compilation process using simple platform and compiler independent configuration files, and generate native makefiles and workspaces that can be used in the compiler environment of your choice.\n\n[GDB](http:\u002F\u002Fwww.gnu.org\u002Fsoftware\u002Fgdb\u002F) is a debugger, that allows you to see what is going on `inside' another program while it executes or what another program was doing at the moment it crashed. \n\n[GCC](https:\u002F\u002Fgcc.gnu.org\u002F) is a compiler Collection that includes front ends for C, C++, Objective-C, Fortran, Ada, Go, and D, as well as libraries for these languages.\n\n[GSL](https:\u002F\u002Fwww.gnu.org\u002Fsoftware\u002Fgsl\u002F) is a numerical library for C and C++ programmers. It is free software under the GNU General Public License. The library provides a wide range of mathematical routines such as random number generators, special functions and least-squares fitting. There are over 1000 functions in total with an extensive test suite.\n\n[OpenGL Extension Wrangler Library (GLEW)](https:\u002F\u002Fwww.opengl.org\u002Fsdk\u002Flibs\u002FGLEW\u002F) is a cross-platform open-source C\u002FC++ extension loading library. GLEW provides efficient run-time mechanisms for determining which OpenGL extensions are supported on the target platform.\n\n[Libtool](https:\u002F\u002Fwww.gnu.org\u002Fsoftware\u002Flibtool\u002F) is a generic library support script that hides the complexity of using shared libraries behind a consistent, portable interface. To use Libtool, add the new generic library building commands to your Makefile, Makefile.in, or Makefile.am.\n\n[Maven](https:\u002F\u002Fmaven.apache.org\u002F) is a software project management and comprehension tool. Based on the concept of a project object model (POM), Maven can manage a project's build, reporting and documentation from a central piece of information.\n\n[TAU (Tuning And Analysis Utilities)](http:\u002F\u002Fwww.cs.uoregon.edu\u002Fresearch\u002Ftau\u002Fhome.php) is capable of gathering performance information through instrumentation of functions, methods, basic blocks, and statements as well as event-based sampling. All C++ language features are supported including templates and namespaces.\n\n[Clang](https:\u002F\u002Fclang.llvm.org\u002F) is a production quality C, Objective-C, C++ and Objective-C++ compiler when targeting X86-32, X86-64, and ARM (other targets may have caveats, but are usually easy to fix). Clang is used in production to build performance-critical software like Google Chrome or Firefox.\n\n[OpenCV](https:\u002F\u002Fopencv.org\u002F) is a highly optimized library with focus on real-time applications. Cross-Platform C++, Python and Java interfaces support Linux, MacOS, Windows, iOS, and Android.\n\n[Libcu++](https:\u002F\u002Fnvidia.github.io\u002Flibcudacxx) is the NVIDIA C++ Standard Library for your entire system. It provides a heterogeneous implementation of the C++ Standard Library that can be used in and between CPU and GPU code.\n\n[ANTLR (ANother Tool for Language Recognition)](https:\u002F\u002Fwww.antlr.org\u002F) is a powerful parser generator for reading, processing, executing, or translating structured text or binary files. It's widely used to build languages, tools, and frameworks. From a grammar, ANTLR generates a parser that can build parse trees and also generates a listener interface that makes it easy to respond to the recognition of phrases of interest.\n\n[Oat++](https:\u002F\u002Foatpp.io\u002F) is a light and powerful C++ web framework for highly scalable and resource-efficient web application. It's zero-dependency and easy-portable.\n\n[JavaCPP](https:\u002F\u002Fgithub.com\u002Fbytedeco\u002Fjavacpp) is a program that provides efficient access to native C++ inside Java, not unlike the way some C\u002FC++ compilers interact with assembly language. \n\n[Cython](https:\u002F\u002Fcython.org\u002F) is a language that makes writing C extensions for Python as easy as Python itself. Cython is based on Pyrex, but supports more cutting edge functionality and optimizations such as calling C functions and declaring C types on variables and class attributes.\n\n[Spdlog](https:\u002F\u002Fgithub.com\u002Fgabime\u002Fspdlog) is a very fast, header-only\u002Fcompiled, C++ logging library. \n\n[Infer](https:\u002F\u002Ffbinfer.com\u002F) is a static analysis tool for Java, C++, Objective-C, and C. Infer is written in [OCaml](https:\u002F\u002Focaml.org\u002F).\n\n# Java Development\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_d898b973591b.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n\n## Java Learning Resources\n\n[Java](https:\u002F\u002Fwww.oracle.com\u002Fjava\u002F) is a popular programming language and development platform(JDK). It reduces costs, shortens development timeframes, drives innovation, and improves application services. With millions of developers running more than 51 billion Java Virtual Machines worldwide.\n\n[The Eclipse Foundation](https:\u002F\u002Fwww.eclipse.org\u002Fdownloads\u002F) is home to a worldwide community of developers, the Eclipse IDE, Jakarta EE and over 375 open source projects, including runtimes, tools and frameworks for Java and other languages.\n\n[Getting Started with Java](https:\u002F\u002Fdocs.oracle.com\u002Fjavase\u002Ftutorial\u002F)\n\n[Oracle Java certifications from Oracle University](https:\u002F\u002Feducation.oracle.com\u002Fjava-certification-benefits)\n\n[Google Developers Training](https:\u002F\u002Fdevelopers.google.com\u002Ftraining\u002F)\n\n[Google Developers Certification](https:\u002F\u002Fdevelopers.google.com\u002Fcertification\u002F)\n\n[Java Tutorial by W3Schools](https:\u002F\u002Fwww.w3schools.com\u002Fjava\u002F)\n\n[Building Your First Android App in Java](codelabs.developers.google.com\u002Fcodelabs\u002Fbuild-your-first-android-app\u002F)\n\n[Getting Started with Java in Visual Studio Code](https:\u002F\u002Fcode.visualstudio.com\u002Fdocs\u002Fjava\u002Fjava-tutorial)\n\n[Google Java Style Guide](https:\u002F\u002Fgoogle.github.io\u002Fstyleguide\u002Fjavaguide.html)\n\n[AOSP Java Code Style for Contributors](https:\u002F\u002Fsource.android.com\u002Fsetup\u002Fcontribute\u002Fcode-style)\n\n[Chromium Java style guide](https:\u002F\u002Fchromium.googlesource.com\u002Fchromium\u002Fsrc\u002F+\u002Fmaster\u002Fstyleguide\u002Fjava\u002Fjava.md)\n\n[Get Started with OR-Tools for Java](https:\u002F\u002Fdevelopers.google.com\u002Foptimization\u002Fintroduction\u002Fjava)\n\n[Getting started with Java Tool Installer task for Azure Pipelines](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fdevops\u002Fpipelines\u002Ftasks\u002Ftool\u002Fjava-tool-installer)\n\n[Gradle User Manual](https:\u002F\u002Fdocs.gradle.org\u002Fcurrent\u002Fuserguide\u002Fuserguide.html)\n\n## Tools\n\n[Java SE](https:\u002F\u002Fwww.oracle.com\u002Fjava\u002Ftechnologies\u002Fjavase\u002Ftools-jsp.html) contains several tools to assist in program development and debugging, and in the monitoring and troubleshooting of production applications. \n\n[JDK Development Tools](https:\u002F\u002Fdocs.oracle.com\u002Fjavase\u002F7\u002Fdocs\u002Ftechnotes\u002Ftools\u002F) includes the Java Web Start Tools (javaws) Java Troubleshooting, Profiling, Monitoring and Management Tools (jcmd, jconsole, jmc, jvisualvm); and Java Web Services Tools (schemagen, wsgen, wsimport, xjc).\n\n[Android Studio](https:\u002F\u002Fdeveloper.android.com\u002Fstudio\u002F) is the official integrated development environment for Google's Android operating system, built on JetBrains' IntelliJ IDEA software and designed specifically for Android development. Availble on Windows, macOS, Linux, Chrome OS.\n\n[IntelliJ IDEA](https:\u002F\u002Fwww.jetbrains.com\u002Fidea\u002F) is an IDE for Java, but it also understands and provides intelligent coding assistance for a large variety of other languages such as Kotlin, SQL, JPQL, HTML, JavaScript, etc., even if the language expression is injected into a String literal in your Java code.\n\n[NetBeans](https:\u002F\u002Fnetbeans.org\u002Ffeatures\u002Fjava\u002Findex.html) is an IDE provides Java developers with all the tools needed to create professional desktop, mobile and enterprise applications. Creating, Editing, and Refactoring. The IDE provides wizards and templates to let you create Java EE, Java SE, and Java ME applications.\n\n[Java Design Patterns ](https:\u002F\u002Fgithub.com\u002Filuwatar\u002Fjava-design-patterns) is a collection of the best formalized practices a programmer can use to solve common problems when designing an application or system.\n\n[Elasticsearch](https:\u002F\u002Fwww.elastic.co\u002Fproducts\u002Felasticsearch) is a distributed RESTful search engine built for the cloud written in Java.\n\n[RxJava](https:\u002F\u002Fgithub.com\u002FReactiveX\u002FRxJava) is a Java VM implementation of [Reactive Extensions](http:\u002F\u002Freactivex.io\u002F): a library for composing asynchronous and event-based programs by using observable sequences. It extends the [observer pattern](http:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FObserver_pattern) to support sequences of data\u002Fevents and adds operators that allow you to compose sequences together declaratively while abstracting away concerns about things like low-level threading, synchronization, thread-safety and concurrent data structures.\n\n[Guava](https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fguava) is a set of core Java libraries from Google that includes new collection types (such as multimap and multiset), immutable collections, a graph library, and utilities for concurrency, I\u002FO, hashing, caching, primitives, strings, and more! It is widely used on most Java projects within Google, and widely used by many other companies as well.\n\n[okhttp](https:\u002F\u002Fsquare.github.io\u002Fokhttp\u002F) is a HTTP client for Java and Kotlin developed by Square. \n\n[Retrofit](https:\u002F\u002Fsquare.github.io\u002Fretrofit\u002F) is a type-safe HTTP client for Android and Java develped by Square.\n\n[LeakCanary](https:\u002F\u002Fsquare.github.io\u002Fleakcanary\u002F) is a memory leak detection library for Android develped by Square.\n\n[Apache Spark](https:\u002F\u002Fspark.apache.org\u002F) is a unified analytics engine for large-scale data processing. It provides high-level APIs in Scala, Java, Python, and R, and an optimized engine that supports general computation graphs for data analysis. It also supports a rich set of higher-level tools including Spark SQL for SQL and DataFrames, MLlib for machine learning, GraphX for graph processing, and Structured Streaming for stream processing.\n\n[Apache Flink](https:\u002F\u002Fflink.apache.org\u002F) is an open source stream processing framework with powerful stream- and batch-processing capabilities with elegant and fluent APIs in Java and Scala.\n\n[Fastjson](https:\u002F\u002Fgithub.com\u002Falibaba\u002Ffastjson\u002Fwiki) is a Java library that can be used to convert Java Objects into their JSON representation. It can also be used to convert a JSON string to an equivalent Java object.\n\n[libGDX](https:\u002F\u002Flibgdx.com\u002F) is a cross-platform Java game development framework based on OpenGL (ES) that works on Windows, Linux, Mac OS X, Android, your WebGL enabled browser and iOS.\n\n[Jenkins](https:\u002F\u002Fwww.jenkins.io\u002F) is the leading open-source automation server. Built with Java, it provides over 1700 [plugins](https:\u002F\u002Fplugins.jenkins.io\u002F) to support automating virtually anything, so that humans can actually spend their time doing things machines cannot.\n\n[DBeaver](https:\u002F\u002Fdbeaver.io\u002F) is a free multi-platform database tool for developers, SQL programmers, database administrators and analysts. Supports any database which has JDBC driver (which basically means - ANY database). EE version also supports non-JDBC datasources (MongoDB, Cassandra, Redis, DynamoDB, etc).\n\n[Redisson](https:\u002F\u002Fredisson.pro\u002F) is a Redis Java client with features of In-Memory Data Grid. Over 50 Redis based Java objects and services: Set, Multimap, SortedSet, Map, List, Queue, Deque, Semaphore, Lock, AtomicLong, Map Reduce, Publish \u002F Subscribe, Bloom filter, Spring Cache, Tomcat, Scheduler, JCache API, Hibernate, MyBatis, RPC, and local cache.\n\n[GraalVM](https:\u002F\u002Fwww.graalvm.org\u002F) is a universal virtual machine for running applications written in JavaScript, Python, Ruby, R, JVM-based languages like Java, Scala, Clojure, Kotlin, and LLVM-based languages such as C and C++.\n\n[Gradle](https:\u002F\u002Fgradle.org\u002F) is a build automation tool for multi-language software development. From mobile apps to microservices, from small startups to big enterprises, Gradle helps teams build, automate and deliver better software, faster. Write in Java, C++, Python or your language of choice. \n\n[Apache Groovy](http:\u002F\u002Fwww.groovy-lang.org\u002F) is a powerful, optionally typed and dynamic language, with static-typing and static compilation capabilities, for the Java platform aimed at improving developer productivity thanks to a concise, familiar and easy to learn syntax. It integrates smoothly with any Java program, and immediately delivers to your application powerful features, including scripting capabilities, Domain-Specific Language authoring, runtime and compile-time meta-programming and functional programming. \n\n[JaCoCo](https:\u002F\u002Fwww.jacoco.org\u002Fjacoco\u002F) is a free code coverage library for Java, which has been created by the EclEmma team based on the lessons learned from using and integration existing libraries for many years.\n\n[Apache JMeter](http:\u002F\u002Fjmeter.apache.org\u002F) is  used to test performance both on static and dynamic resources, Web dynamic applications. It also used to simulate a heavy load on a server, group of servers, network or object to test its strength or to analyze overall performance under different load types.\n\n[Junit](https:\u002F\u002Fjunit.org\u002F) is a simple framework to write repeatable tests. It is an instance of the xUnit architecture for unit testing frameworks.\n\n[Mockito](https:\u002F\u002Fsite.mockito.org\u002F) is the most popular Mocking framework for unit tests written in Java.\n\n[SpotBugs](https:\u002F\u002Fspotbugs.github.io\u002F) is a program which uses static analysis to look for bugs in Java code.\n\n[SpringBoot](https:\u002F\u002Fspring.io\u002Fprojects\u002Fspring-boot) is a great tool that helps you to create Spring-powered, production-grade applications and services with absolute minimum fuss. It takes an opinionated view of the Spring platform so that new and existing users can quickly get to the bits they need.\n\n[YourKit](https:\u002F\u002Fwww.yourkit.com\u002F) is a technology leader, creator of the most innovative and intelligent tools for profiling Java & .NET applications.\n\n# Python Development\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_83a9b576f0a0.png\">\n  \u003Cbr \u002F>\n\n\u003C\u002Fp>\n\n## Python Learning Resources\n\n[Python](https:\u002F\u002Fwww.python.org) is an interpreted, high-level programming language. Python is used heavily in the fields of Data Science and Machine Learning. \n\n[Python Developer’s Guide](https:\u002F\u002Fdevguide.python.org) is a comprehensive resource for contributing to Python – for both new and experienced contributors. It is maintained by the same community that maintains Python. \n\n[Azure Functions Python developer guide](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fazure-functions\u002Ffunctions-reference-python) is an introduction to developing Azure Functions using Python. The content below assumes that you've already read the [Azure Functions developers guide](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fazure-functions\u002Ffunctions-reference).\n\n[CheckiO](https:\u002F\u002Fcheckio.org\u002F) is a programming learning platform and a gamified website that teaches Python through solving code challenges and competing for the most elegant and creative solutions.\n\n[Python Institute](https:\u002F\u002Fpythoninstitute.org)\n\n[PCEP – Certified Entry-Level Python Programmer certification](https:\u002F\u002Fpythoninstitute.org\u002Fpcep-certification-entry-level\u002F)\n\n[PCAP – Certified Associate in Python Programming certification](https:\u002F\u002Fpythoninstitute.org\u002Fpcap-certification-associate\u002F)\n\n[PCPP – Certified Professional in Python Programming 1 certification](https:\u002F\u002Fpythoninstitute.org\u002Fpcpp-certification-professional\u002F)\n\n[PCPP – Certified Professional in Python Programming 2](https:\u002F\u002Fpythoninstitute.org\u002Fpcpp-certification-professional\u002F)\n\n[MTA: Introduction to Programming Using Python Certification](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Flearn\u002Fcertifications\u002Fmta-introduction-to-programming-using-python)\n\n[Getting Started with Python in Visual Studio Code](https:\u002F\u002Fcode.visualstudio.com\u002Fdocs\u002Fpython\u002Fpython-tutorial)\n\n[Google's Python Style Guide](https:\u002F\u002Fgoogle.github.io\u002Fstyleguide\u002Fpyguide.html)\n\n[Google's Python Education Class](https:\u002F\u002Fdevelopers.google.com\u002Fedu\u002Fpython\u002F)\n\n[Real Python](https:\u002F\u002Frealpython.com)\n\n[The Python Open Source Computer Science Degree by Forrest Knight](https:\u002F\u002Fgithub.com\u002FForrestKnight\u002Fopen-source-cs-python)\n\n[Intro to Python for Data Science](https:\u002F\u002Fwww.datacamp.com\u002Fcourses\u002Fintro-to-python-for-data-science)\n\n[Intro to Python by W3schools](https:\u002F\u002Fwww.w3schools.com\u002Fpython\u002Fpython_intro.asp)\n\n[Codecademy's Python 3 course](https:\u002F\u002Fwww.codecademy.com\u002Flearn\u002Flearn-python-3)\n\n[Learn Python with Online Courses and Classes from edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fpython)\n\n[Python Courses Online from Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=python)\n\n## Python Frameworks and Tools\n\n[Python Package Index (PyPI)](https:\u002F\u002Fpypi.org\u002F) is a repository of software for the Python programming language. PyPI helps you find and install software developed and shared by the Python community. \n\n[PyCharm](https:\u002F\u002Fwww.jetbrains.com\u002Fpycharm\u002F) is the best IDE I've ever used. With PyCharm, you can access the command line, connect to a database, create a virtual environment, and manage your version control system all in one place, saving time by avoiding constantly switching between windows.\n\n[Python Tools for Visual Studio(PTVS)](https:\u002F\u002Fmicrosoft.github.io\u002FPTVS\u002F) is a free, open source plugin that turns Visual Studio into a Python IDE. It supports editing, browsing, IntelliSense, mixed Python\u002FC++ debugging, remote Linux\u002FMacOS debugging, profiling, IPython, and web development with Django and other frameworks.\n\n[Pylance](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fpylance-release) is an extension that works alongside Python in Visual Studio Code to provide performant language support. Under the hood, Pylance is powered by Pyright, Microsoft's static type checking tool.\n\n[Pyright](https:\u002F\u002Fgithub.com\u002FMicrosoft\u002Fpyright) is a fast type checker meant for large Python source bases. It can run in a “watch” mode and performs fast incremental updates when files are modified.\n\n[Django](https:\u002F\u002Fwww.djangoproject.com\u002F) is a high-level Python Web framework that encourages rapid development and clean, pragmatic design.\n\n[Flask](https:\u002F\u002Fflask.palletsprojects.com\u002F) is a micro web framework written in Python. It is classified as a microframework because it does not require particular tools or libraries. \n \n[Web2py](http:\u002F\u002Fweb2py.com\u002F) is an open-source web application framework written in Python allowing allows web developers to program dynamic web content. One web2py instance can run multiple web sites using different databases.\n\n[AWS Chalice](https:\u002F\u002Fgithub.com\u002Faws\u002Fchalice) is a framework for writing serverless apps in python. It allows you to quickly create and deploy applications that use AWS Lambda. \n\n[Tornado](https:\u002F\u002Fwww.tornadoweb.org\u002F) is a Python web framework and asynchronous networking library. Tornado uses a non-blocking network I\u002FO, which can scale to tens of thousands of open connections.\n\n[HTTPie](https:\u002F\u002Fgithub.com\u002Fhttpie\u002Fhttpie) is a command line HTTP client that makes CLI interaction with web services as easy as possible. HTTPie is designed for testing, debugging, and generally interacting with APIs & HTTP servers. \n\n[Scrapy](https:\u002F\u002Fscrapy.org\u002F) is a fast high-level web crawling and web scraping framework, used to crawl websites and extract structured data from their pages. It can be used for a wide range of purposes, from data mining to monitoring and automated testing.\n\n[Sentry](https:\u002F\u002Fsentry.io\u002F) is a service that helps you monitor and fix crashes in realtime. The server is in Python, but it contains a full API for sending events from any language, in any application.\n\n[Pipenv](https:\u002F\u002Fgithub.com\u002Fpypa\u002Fpipenv) is a tool that aims to bring the best of all packaging worlds (bundler, composer, npm, cargo, yarn, etc.) to the Python world.\n\n[Python Fire](https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fpython-fire) is a library for automatically generating command line interfaces (CLIs) from absolutely any Python object.\n\n[Bottle](https:\u002F\u002Fgithub.com\u002Fbottlepy\u002Fbottle) is a fast, simple and lightweight [WSGI](https:\u002F\u002Fwww.wsgi.org\u002F) micro web-framework for Python. It is distributed as a single file module and has no dependencies other than the [Python Standard Library](https:\u002F\u002Fdocs.python.org\u002Flibrary\u002F).\n\n[CherryPy](https:\u002F\u002Fcherrypy.org) is a minimalist Python object-oriented HTTP web framework.\n\n[Sanic](https:\u002F\u002Fgithub.com\u002Fhuge-success\u002Fsanic) is a Python 3.6+ web server and web framework that's written to go fast. \n\n[Pyramid](https:\u002F\u002Ftrypyramid.com) is a small and fast open source Python web framework. It makes real-world web application development and deployment more fun and more productive.\n\n[TurboGears](https:\u002F\u002Fturbogears.org) is a hybrid web framework able to act both as a Full Stack framework or as a Microframework. \n\n[Falcon](https:\u002F\u002Ffalconframework.org\u002F) is a reliable, high-performance Python web framework for building large-scale app backends and microservices with support for MongoDB, Pluggable Applications and autogenerated Admin.\n\n[Neural Network Intelligence(NNI)](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fnni) is an open source AutoML toolkit for automate machine learning lifecycle, including [Feature Engineering](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fnni\u002Fblob\u002Fmaster\u002Fdocs\u002Fen_US\u002FFeatureEngineering\u002FOverview.md), [Neural Architecture Search](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fnni\u002Fblob\u002Fmaster\u002Fdocs\u002Fen_US\u002FNAS\u002FOverview.md), [Model Compression](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fnni\u002Fblob\u002Fmaster\u002Fdocs\u002Fen_US\u002FCompressor\u002FOverview.md) and [Hyperparameter Tuning](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fnni\u002Fblob\u002Fmaster\u002Fdocs\u002Fen_US\u002FTuner\u002FBuiltinTuner.md).\n\n[Dash](https:\u002F\u002Fplotly.com\u002Fdash) is a popular Python framework for building ML & data science web apps for Python, R, Julia, and Jupyter.\n\n[Luigi](https:\u002F\u002Fgithub.com\u002Fspotify\u002Fluigi) is a Python module that helps you build complex pipelines of batch jobs. It handles dependency resolution, workflow management, visualization etc. It also comes with Hadoop support built-in.\n\n[Locust](https:\u002F\u002Fgithub.com\u002Flocustio\u002Flocust) is an easy to use, scriptable and scalable performance testing tool. \n\n[spaCy](https:\u002F\u002Fgithub.com\u002Fexplosion\u002FspaCy) is a library for advanced Natural Language Processing in Python and Cython. \n\n[NumPy](https:\u002F\u002Fwww.numpy.org\u002F) is the fundamental package needed for scientific computing with Python.\n\n[Pillow](https:\u002F\u002Fpython-pillow.org\u002F) is a friendly PIL(Python Imaging Library) fork.\n\n[IPython](https:\u002F\u002Fipython.org\u002F) is a command shell for interactive computing in multiple programming languages, originally developed for the Python programming language, that offers enhanced introspection, rich media, additional shell syntax, tab completion, and rich history.\n\n[GraphLab Create](https:\u002F\u002Fturi.com\u002F) is a Python library, backed by a C++ engine, for quickly building large-scale, high-performance machine learning models.\n\n[Pandas](https:\u002F\u002Fpandas.pydata.org\u002F) is a fast, powerful, and easy to use open source data structrures, data analysis and manipulation tool, built on top of the Python programming language.\n\n[PuLP](https:\u002F\u002Fcoin-or.github.io\u002Fpulp\u002F) is an Linear Programming modeler written in python. PuLP can generate LP files and call on use highly optimized solvers, GLPK, COIN CLP\u002FCBC, CPLEX, and GUROBI, to solve these linear problems.\n\n[Matplotlib](https:\u002F\u002Fmatplotlib.org\u002F) is a 2D plotting library for creating static, animated, and interactive visualizations in Python. Matplotlib produces publication-quality figures in a variety of hardcopy formats and interactive environments across platforms.\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) is a simple and efficient tool for data mining and data analysis. It is built on NumPy,SciPy, and mathplotlib.\n\n\n# Scala Development\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_334a63768e64.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n\n# Scala Learning Resources\n\n[Scala](https:\u002F\u002Fscala-lang.org\u002F) is a combination of object-oriented and functional programming in one concise, high-level language. Scala's static types help avoid bugs in complex applications, and its JVM and JavaScript runtimes let you build high-performance systems with easy access to huge ecosystems of libraries.\n\n[Scala Style Guide](https:\u002F\u002Fdocs.scala-lang.org\u002Fstyle\u002F)\n\n[Databricks Scala Style Guide](https:\u002F\u002Fgithub.com\u002Fdatabricks\u002Fscala-style-guide)\n\n[Data Science using Scala and Spark on Azure](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fmachine-learning\u002Fteam-data-science-process\u002Fscala-walkthrough)\n\n[Creating a Scala Maven application for Apache Spark in HDInsight using IntelliJ](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fhdinsight\u002Fspark\u002Fapache-spark-create-standalone-application)\n\n[Intro to Spark DataFrames using Scala with Azure Databricks](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fdatabricks\u002Fspark\u002Flatest\u002Fdataframes-datasets\u002Fintroduction-to-dataframes-scala)\n\n[Using Scala to Program AWS Glue ETL Scripts](https:\u002F\u002Fdocs.aws.amazon.com\u002Fglue\u002Flatest\u002Fdg\u002Fglue-etl-scala-using.html)\n\n[Using Flink Scala shell with Amazon EMR clusters](https:\u002F\u002Fdocs.aws.amazon.com\u002Femr\u002Flatest\u002FReleaseGuide\u002Fflink-scala.html)\n\n[AWS EMR and Spark 2 using Scala from Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Faws-emr-and-spark-2-using-scala\u002F)\n\n[Using the Google Cloud Storage connector with Apache Spark](https:\u002F\u002Fcloud.google.com\u002Fdataproc\u002Fdocs\u002Ftutorials\u002Fgcs-connector-spark-tutorial)\n\n[Write and run Spark Scala jobs on Cloud Dataproc for Google Cloud](https:\u002F\u002Fcloud.google.com\u002Fdataproc\u002Fdocs\u002Ftutorials\u002Fspark-scala)\n\n[Scala Courses and Certifications from edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fscala)\n\n[Scala Courses from Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=scala)\n\n[Top Scala Courses from Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fscala\u002F)\n\n# Scala Tools\n\n[Apache Spark](https:\u002F\u002Fspark.apache.org\u002F) is a unified analytics engine for large-scale data processing. It provides high-level APIs in Scala, Java, Python, and R, and an optimized engine that supports general computation graphs for data analysis. It also supports a rich set of higher-level tools including Spark SQL for SQL and DataFrames, MLlib for machine learning, GraphX for graph processing, and Structured Streaming for stream processing.\n\n[Apache Spark Connector for SQL Server and Azure SQL](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fsql-spark-connector) is a high-performance connector that enables you to use transactional data in big data analytics and persists results for ad-hoc queries or reporting. The connector allows you to use any SQL database, on-premises or in the cloud, as an input data source or output data sink for Spark jobs.\n\n[Azure Databricks](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fdatabricks\u002F) is a fast and collaborative Apache Spark-based big data analytics service designed for data science and data engineering. Azure Databricks, sets up your Apache Spark environment in minutes, autoscale, and collaborate on shared projects in an interactive workspace. Azure Databricks supports Python, Scala, R, Java, and SQL, as well as data science frameworks and libraries including TensorFlow, PyTorch, and scikit-learn.\n\n[Apache PredictionIO](https:\u002F\u002Fpredictionio.apache.org\u002F) is an open source machine learning framework for developers, data scientists, and end users. It supports event collection, deployment of algorithms, evaluation, querying predictive results via REST APIs. It is based on scalable open source services like Hadoop, HBase (and other DBs), Elasticsearch, Spark and implements what is called a Lambda Architecture.\n\n[Cluster Manager for Apache Kafka(CMAK)](https:\u002F\u002Fgithub.com\u002Fyahoo\u002FCMAK) is a tool for managing [Apache Kafka](https:\u002F\u002Fkafka.apache.org\u002F) clusters.\n\n[BigDL](https:\u002F\u002Fbigdl-project.github.io\u002F) is a distributed deep learning library for Apache Spark. With BigDL, users can write their deep learning applications as standard Spark programs, which can directly run on top of existing Spark or Hadoop clusters.\n\n[Eclipse Deeplearning4J (DL4J)](https:\u002F\u002Fdeeplearning4j.konduit.ai\u002F) is a set of projects intended to support all the needs of a JVM-based(Scala, Kotlin, Clojure, and Groovy) deep learning application. This means starting with the raw data, loading and preprocessing it from wherever and whatever format it is in to building and tuning a wide variety of simple and complex deep learning networks.\n\n[Play Framework](https:\u002F\u002Fgithub.com\u002Fplayframework\u002Fplayframework) is a web framework combines productivity and performance making it easy to build scalable web applications with Java and Scala. \n\n[Dotty](https:\u002F\u002Fgithub.com\u002Flampepfl\u002Fdotty) is a research compiler that will become Scala 3.\n\n[AWScala](https:\u002F\u002Fgithub.com\u002Fseratch\u002FAWScala) is a tool that enables Scala developers to easily work with Amazon Web Services in the Scala way.\n\n[Scala.js](https:\u002F\u002Fwww.scala-js.org\u002F) is a compiler that converts Scala to JavaScript.\n\n[Polynote](https:\u002F\u002Fpolynote.org\u002F) is an experimental polyglot notebook environment. Currently, it supports Scala and Python (with or without Spark), SQL, and Vega.\n\n[Scala Native](http:\u002F\u002Fscala-native.org\u002F) is an optimizing ahead-of-time compiler and lightweight managed runtime designed specifically for Scala. \n\n[Gitbucket](https:\u002F\u002Fgitbucket.github.io\u002F) is a Git platform powered by Scala with easy installation, high extensibility & GitHub API compatibility.\n\n[Finagle](https:\u002F\u002Ftwitter.github.io\u002Ffinagle) is a fault tolerant, protocol-agnostic RPC system\n\n[Gatling](https:\u002F\u002Fgatling.io\u002F) is a load test tool. It officially supports HTTP, WebSocket, Server-Sent-Events and JMS.\n\n[Scalatra](https:\u002F\u002Fscalatra.org\u002F) is a tiny Scala high-performance, async web framework, inspired by [Sinatra](https:\u002F\u002Fwww.sinatrarb.com\u002F).\n\n\n# R Development\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_e3cb2c3b58cd.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n# R Learning Resources\n\n[R](https:\u002F\u002Fwww.r-project.org\u002F) is an open source software environment for statistical computing and graphics. It compiles and runs on a wide variety of  platforms such as Windows and MacOS. \n\n[An Introduction to R](https:\u002F\u002Fcran.r-project.org\u002Fdoc\u002Fmanuals\u002Fr-release\u002FR-intro.pdf)\n\n[Google's R Style Guide](https:\u002F\u002Fgoogle.github.io\u002Fstyleguide\u002FRguide.html)\n\n[R developer's guide to Azure](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Farchitecture\u002Fdata-guide\u002Ftechnology-choices\u002Fr-developers-guide)\n\n[Running R at Scale on Google Compute Engine](https:\u002F\u002Fcloud.google.com\u002Fsolutions\u002Frunning-r-at-scale)\n\n[Running R on AWS](https:\u002F\u002Faws.amazon.com\u002Fblogs\u002Fbig-data\u002Frunning-r-on-aws\u002F)\n\n[RStudio Server Pro for AWS](https:\u002F\u002Faws.amazon.com\u002Fmarketplace\u002Fpp\u002FRStudio-RStudio-Server-Pro-for-AWS\u002FB06W2G9PRY)\n\n[Learn R by Codecademy](https:\u002F\u002Fwww.codecademy.com\u002Flearn\u002Flearn-r)\n\n[Learn R Programming with Online Courses and Lessons by edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fr-programming)\n\n[R Language Courses by Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=r%20language)\n\n[Learn R For Data Science by Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fprogramming-for-data-science-nanodegree-with-R--nd118)\n\n# R Tools\n\n[RStudio](https:\u002F\u002Frstudio.com\u002F) is an integrated development environment for R and Python, with a console, syntax-highlighting editor that supports direct code execution, and tools for plotting, history, debugging and workspace management.\n\n[Shiny](https:\u002F\u002Fshiny.rstudio.com\u002F) is a newer package from RStudio that makes it incredibly easy to build interactive web applications with R.\n\n[Rmarkdown ](https:\u002F\u002Frmarkdown.rstudio.com\u002F) is a package helps you create dynamic analysis documents that combine code, rendered output (such as figures), and prose. \n\n[Rplugin](https:\u002F\u002Fgithub.com\u002FJetBrains\u002FRplugin) is R Language supported plugin for the IntelliJ IDE.\n\n[Plotly](https:\u002F\u002Fplotly-r.com\u002F) is an R package for creating interactive web graphics via the open source JavaScript graphing library [plotly.js](https:\u002F\u002Fgithub.com\u002Fplotly\u002Fplotly.js).\n\n[Metaflow](https:\u002F\u002Fmetaflow.org\u002F) is a Python\u002FR library that helps scientists and engineers build and manage real-life data science projects. Metaflow was originally developed at Netflix to boost productivity of data scientists who work on a wide variety of projects from classical statistics to state-of-the-art deep learning.\n\n[Prophet](https:\u002F\u002Ffacebook.github.io\u002Fprophet) is a procedure for forecasting time series data based on an additive model where non-linear trends are fit with yearly, weekly, and daily seasonality, plus holiday effects. It works best with time series that have strong seasonal effects and several seasons of historical data.\n\n[LightGBM](https:\u002F\u002Flightgbm.readthedocs.io\u002F) is a gradient boosting framework that uses tree based learning algorithms, used for ranking, classification and many other machine learning tasks. \n\n[Dash](https:\u002F\u002Fplotly.com\u002Fdash) is a Python framework for building analytical web applications in Python, R, Julia, and Jupyter.\n\n[MLR](https:\u002F\u002Fmlr.mlr-org.com\u002F) is Machine Learning in R.\n\n[ML workspace](https:\u002F\u002Fgithub.com\u002Fml-tooling\u002Fml-workspace) is an all-in-one web-based IDE specialized for machine learning and data science. It is simple to deploy and gets you started within minutes to productively built ML solutions on your own machines. ML workspace is the ultimate tool for developers preloaded with a variety of popular data science libraries (Tensorflow, PyTorch, Keras, and MXnet) and dev tools (Jupyter, VS Code, and Tensorboard) perfectly configured, optimized, and integrated.\n\n[CatBoost](https:\u002F\u002Fcatboost.ai\u002F) is a fast, scalable, high performance Gradient Boosting on Decision Trees library, used for ranking, classification, regression and other machine learning tasks for Python, R, Java, C++. Supports computation on CPU and GPU.\n\n[Plumber](https:\u002F\u002Fwww.rplumber.io\u002F) is a tool that allows you to create a web API by merely decorating your existing R source code with special comments. \n\n[Drake](https:\u002F\u002Fdocs.ropensci.org\u002Fdrake) is an R-focused pipeline toolkit for reproducibility and high-performance computing.\n\n[DiagrammeR](https:\u002F\u002Fvisualizers.co\u002Fdiagrammer\u002F) is a package you can create, modify, analyze, and visualize network graph diagrams. The output can be incorporated into R Markdown documents, integrated with Shiny web apps, converted to other graph formats, or exported as image files.\n\n[Knitr](https:\u002F\u002Fyihui.org\u002Fknitr\u002F) is a general-purpose literate programming engine in R, with lightweight API's designed to give users full control of the output without heavy coding work.\n\n[Broom](https:\u002F\u002Fbroom.tidymodels.org\u002F) is a tool that converts statistical analysis objects from R into tidy format.\n\n\n# Julia Development\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_1424f02ea237.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n# Julia Learning Resources\n\n[Julia](https:\u002F\u002Fjulialang.org) is a high-level, [high-performance](https:\u002F\u002Fjulialang.org\u002Fbenchmarks\u002F) dynamic language for technical computing. Julia programs compile to efficient native code for [multiple platforms](https:\u002F\u002Fjulialang.org\u002Fdownloads\u002F#support_tiers) via LLVM.\n\n[JuliaHub](https:\u002F\u002Fjuliahub.com\u002F) contains over 4,000 Julia packages for use by the community.\n\n[Julia Observer](https:\u002F\u002Fwww.juliaobserver.com)\n\n[Julia Manual](https:\u002F\u002Fdocs.julialang.org\u002Fen\u002Fv1\u002Fmanual\u002Fgetting-started\u002F)\n\n[JuliaLang Essentials](https:\u002F\u002Fdocs.julialang.org\u002Fen\u002Fv1\u002Fbase\u002Fbase\u002F)\n\n[Julia Style Guide](https:\u002F\u002Fdocs.julialang.org\u002Fen\u002Fv1\u002Fmanual\u002Fstyle-guide\u002F)\n\n[Julia By Example](https:\u002F\u002Fjuliabyexample.helpmanual.io\u002F)\n\n[JuliaLang Gitter](https:\u002F\u002Fgitter.im\u002FJuliaLang\u002Fjulia)\n\n[DataFrames Tutorial using Jupyter Notebooks](https:\u002F\u002Fgithub.com\u002Fbkamins\u002FJulia-DataFrames-Tutorial\u002F)\n\n[Julia Academy](https:\u002F\u002Fjuliaacademy.com\u002Fcourses?preview=logged_out)\n\n[Julia Meetup groups](https:\u002F\u002Fwww.meetup.com\u002Ftopics\u002Fjulia\u002F)\n\n[Julia on Microsoft Azure](https:\u002F\u002Fjuliacomputing.com\u002Fmedia\u002F2017\u002F02\u002F08\u002Fazure.html)\n\n# Julia Tools\n\n[JuliaPro](https:\u002F\u002Fjuliacomputing.com\u002Fproducts\u002Fjuliapro.html) is a free and fast way to setup Julia for individual researchers, engineers, scientists, quants, traders, economists, students and others. Julia developers can build better software quicker and easier while benefiting from Julia's unparalleled high performance. It includes 2600+ open source packages or from a curated list of 250+ JuliaPro packages. Curated packages are tested, documented and supported by Julia Computing.\n\n[Juno](https:\u002F\u002Fjunolab.org) is a powerful, free IDE based on [Atom]() for the Julia language.\n\n[Debugger.jl](https:\u002F\u002Fgithub.com\u002FJuliaDebug\u002FDebugger.jl) is the Julia debuggin tool.\n\n[Profile (Stdlib)](https:\u002F\u002Fdocs.julialang.org\u002Fen\u002Fv1\u002Fmanual\u002Fprofile\u002F) is a module provides tools to help developers improve the performance of their code. When used, it takes measurements on running code, and produces output that helps you understand how much time is spent on individual line's.\n\n[Revise.jl](https:\u002F\u002Fgithub.com\u002Ftimholy\u002FRevise.jl) allows you to modify code and use the changes without restarting Julia. With Revise, you can be in the middle of a session and then update packages, switch git branches, and\u002For edit the source code in the editor of your choice; any changes will typically be incorporated into the very next command you issue from the REPL. This can save you the overhead of restarting Julia, loading packages, and waiting for code to JIT-compile.\n\n[JuliaGPU](https:\u002F\u002Fjuliagpu.org\u002F) is a Github organization created to unify the many packages for programming GPUs in Julia. With its high-level syntax and flexible compiler, Julia is well positioned to productively program hardware accelerators like GPUs without sacrificing performance.\n\n[IJulia.jl](https:\u002F\u002Fgithub.com\u002FJuliaLang\u002FIJulia.jl) is the Julia kernel for Jupyter.\n\n[AWS.jl](https:\u002F\u002Fgithub.com\u002FJuliaCloud\u002FAWS.jl) is a Julia interface for [Amazon Web Services](https:\u002F\u002Faws.amazon.com\u002F).\n\n[CUDA.jl](https:\u002F\u002Fjuliagpu.gitlab.io\u002FCUDA.jl) is a package for the main programming interface for working with NVIDIA CUDA GPUs using Julia. It features a user-friendly array abstraction, a compiler for writing CUDA kernels in Julia, and wrappers for various CUDA libraries.\n\n[XLA.jl](https:\u002F\u002Fgithub.com\u002FJuliaTPU\u002FXLA.jl) is a package for compiling Julia to XLA for [Tensor Processing Unit(TPU)](https:\u002F\u002Fcloud.google.com\u002Ftpu\u002F).\n\n[Nanosoldier.jl](https:\u002F\u002Fgithub.com\u002FJuliaCI\u002FNanosoldier.jl) is a package for running JuliaCI services on MIT's Nanosoldier cluster.\n\n[Julia for VSCode](https:\u002F\u002Fwww.julia-vscode.org) is a powerful extension for the Julia language.\n\n[JuMP.jl](https:\u002F\u002Fjump.dev\u002F) is a domain-specific modeling language for [mathematical optimization](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FMathematical_optimization) embedded in Julia.\n\n[Optim.jl](https:\u002F\u002Fgithub.com\u002FJuliaNLSolvers\u002FOptim.jl) is a univariate and multivariate optimization in Julia.\n\n[RCall.jl](https:\u002F\u002Fgithub.com\u002FJuliaInterop\u002FRCall.jl) is a package that allows you to call R functions from Julia.\n\n[JavaCall.jl](http:\u002F\u002Fjuliainterop.github.io\u002FJavaCall.jl) is a package that allows you to call Java functions from Julia.\n\n[PyCall.jl](https:\u002F\u002Fgithub.com\u002FJuliaPy\u002FPyCall.jl) is a package that allows you to call Python functions from Julia.\n\n[MXNet.jl](https:\u002F\u002Fgithub.com\u002Fdmlc\u002FMXNet.jl) is the Apache MXNet Julia package. MXNet.jl brings flexible and efficient GPU computing and state-of-art deep learning to Julia.\n\n[Knet](https:\u002F\u002Fdenizyuret.github.io\u002FKnet.jl\u002Flatest) is the [Koç University deep](http:\u002F\u002Fwww.ku.edu.tr\u002Fen) learning framework implemented in Julia by [Deniz Yuret](https:\u002F\u002Fwww.denizyuret.com\u002F) and collaborators. It supports GPU operation and automatic differentiation using dynamic computational graphs for models defined in plain Julia.\n\n[Distributions.jl](https:\u002F\u002Fgithub.com\u002FJuliaStats\u002FDistributions.jl) is a Julia package for probability distributions and associated functions. \n\n[DataFrames.jl](http:\u002F\u002Fjuliadata.github.io\u002FDataFrames.jl\u002Fstable\u002F) is a tool for working with tabular data in Julia.\n\n[Flux.jl](https:\u002F\u002Ffluxml.ai\u002F) is an elegant approach to machine learning. It's a 100% pure-Julia stack, and provides lightweight abstractions on top of Julia's native GPU and AD support.\n\n[IRTools.jl](https:\u002F\u002Fgithub.com\u002FFluxML\u002FIRTools.jl) is a simple and flexible IR format, expressive enough to work with both lowered and typed Julia code, as well as external IRs.\n\n[Cassette.jl](https:\u002F\u002Fgithub.com\u002Fjrevels\u002FCassette.jl) is a Julia package that provides a mechanism for dynamically injecting code transformation passes into Julia’s just-in-time (JIT) compilation cycle, enabling post hoc analysis and modification of \"Cassette-unaware\" Julia programs without requiring manual source annotation or refactoring of the target code.\n\n## Contribute\n\n- [x] If would you like to contribute to this guide simply make a [Pull Request](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide\u002Fpulls).\n\n\n## License\n\n[Back to the Top](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\nDistributed under the [Creative Commons Attribution 4.0 International (CC BY 4.0) Public License](https:\u002F\u002Fcreativecommons.org\u002Flicenses\u002Fby\u002F4.0\u002F).\n","\u003Ch1 align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_fa28eb9b69ce.png\">\n  \u003Cbr \u002F>\n  机器学习指南\n\u003C\u002Fh1>\n\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fmikeroyal?tab=followers\">\n         \u003Cimg alt=\"followers\" title=\"关注我获取最新动态\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_e273f3008dc5.png\"\u002F>\u003C\u002Fa> \t\n\n![维护中](https:\u002F\u002Fimg.shields.io\u002Fmaintenance\u002Fyes\u002F2024?style=for-the-badge)\n![最近提交](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flast-commit\u002Fmikeroyal\u002Fmachine-learning-guide?style=for-the-badge)\n\n#### 本指南涵盖机器学习的各个方面，包括应用场景、常用库和工具，助您更高效地进行机器学习开发。\n\n **注：您可以在 [VSCode](https:\u002F\u002Fcode.visualstudio.com\u002F) 中使用便捷扩展 [Markdown PDF](https:\u002F\u002Fmarketplace.visualstudio.com\u002Fitems?itemName=yzane.markdown-pdf) 轻松将此 Markdown 文件转换为 PDF。**\n \n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_955586009a84.png\">\n\n**机器学习\u002F深度学习框架。**\n\n# 目录\n\n1. [机器学习学习资源](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#learning-resources-for-ML)\n\n     - [开发者资源](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#developer-resources)\n     - [课程与认证](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#courses--certifications)\n     - [书籍](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#books)\n     - [YouTube 教程](#youtube-tutorials)\n\n2. [机器学习框架、库和工具](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#ML-frameworks-libraries-and-tools)\n\n    - [LLM 训练框架](#llm-training-frameworks)\n    - [部署 LLM 的工具](#tools-for-deploying-llm)\n    - [本地运行大型语言模型 (LLMs)](#running-llms-locally)\n    \n3. [算法](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#Algorithms)\n\n4. [PyTorch 开发](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#pytorch-development)\n\n5. [TensorFlow 开发](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#tensorflow-development)\n\n6. [Core ML 开发](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#core-ml-development)\n\n7. [深度学习开发](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#Deep-Learning-Development)\n\n8. [强化学习开发](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#Reinforcement-Learning-Development)\n\n9. [计算机视觉开发](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#computer-vision-development)\n\n10. [自然语言处理 (NLP) 开发](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#nlp-development)\n\n11. [生物信息学](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#bioinformatics)\n\n12. [CUDA 开发](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#cuda-development)\n\n13. [MATLAB 开发](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#matlab-development)\n\n14. [C\u002FC++ 开发](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#cc-development)\n\n15. [Java 开发](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#java-development)\n\n16. [Python 开发](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#python-development)\n\n17. [Scala 开发](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#scala-development)\n\n18. [R 开发](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#r-development)\n\n19. [Julia 开发](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#julia-development)\n\n\n# 机器学习学习资源\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n[机器学习](https:\u002F\u002Fwww.ibm.com\u002Fcloud\u002Flearn\u002Fmachine-learning) 是人工智能 (AI) 的一个分支，专注于利用从数据模型中学习并随着时间推移不断提高准确性的算法来构建应用程序，而无需显式编程。\n\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_95dd8ee4e1dd.jpeg\">\n\n### 开发者资源\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n- [微软的自然语言处理 (NLP) 最佳实践](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fnlp-recipes)\n\n- [微软的自动驾驶 Cookbook](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002FAutonomousDrivingCookbook)\n\n- [Azure 机器学习 - 作为服务的机器学习 | Microsoft Azure](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fmachine-learning\u002F)\n\n- [如何在 Azure 机器学习工作区中运行 Jupyter 笔记本](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fmachine-learning\u002Fhow-to-run-jupyter-notebooks)\n\n- [机器学习和人工智能 | Amazon Web Services](https:\u002F\u002Faws.amazon.com\u002Fmachine-learning\u002F)\n\n- [在 Amazon SageMaker 短暂实例上调度 Jupyter 笔记本](https:\u002F\u002Faws.amazon.com\u002Fblogs\u002Fmachine-learning\u002Fscheduling-jupyter-notebooks-on-sagemaker-ephemeral-instances\u002F)\n\n- [AI 和机器学习 | Google Cloud](https:\u002F\u002Fcloud.google.com\u002Fproducts\u002Fai\u002F)\n\n- [在 Google Cloud 上使用 Apache Spark 运行 Jupyter 笔记本](https:\u002F\u002Fcloud.google.com\u002Fblog\u002Fproducts\u002Fgcp\u002Fgoogle-cloud-platform-for-data-scientists-using-jupyter-notebooks-with-apache-spark-on-google-cloud)\n\n- [机器学习 | Apple 开发者](https:\u002F\u002Fdeveloper.apple.com\u002Fmachine-learning\u002F)\n\n- [人工智能和自动驾驶 | Tesla](https:\u002F\u002Fwww.tesla.com\u002FAI)\n\n- [Meta AI 工具 | Facebook](https:\u002F\u002Fai.facebook.com\u002Ftools\u002F)\n\n- [PyTorch 教程](https:\u002F\u002Fpytorch.org\u002Ftutorials\u002F)\n\n- [TensorFlow 教程](https:\u002F\u002Fwww.tensorflow.org\u002Ftutorials)\n\n- [JupyterLab](https:\u002F\u002Fjupyterlab.readthedocs.io\u002F)\n\n- [在 Apple Silicon 上使用 Core ML 运行 Stable Diffusion](https:\u002F\u002Fmachinelearning.apple.com\u002Fresearch\u002Fstable-diffusion-coreml-apple-silicon)\n\n### 课程与认证\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n- [斯坦福大学Andrew Ng主讲的机器学习课程 | Coursera](https:\u002F\u002Fwww.coursera.org\u002Flearn\u002Fmachine-learning)\n\n- [AWS机器学习（ML）培训与认证课程](https:\u002F\u002Faws.amazon.com\u002Ftraining\u002Flearning-paths\u002Fmachine-learning\u002F)\n\n- [微软Azure机器学习奖学金项目 | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fscholarships\u002Fmachine-learning-scholarship-microsoft-azure)\n\n- [微软认证：Azure数据科学家助理](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Flearn\u002Fcertifications\u002Fazure-data-scientist)\n\n- [微软认证：Azure人工智能工程师助理](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Flearn\u002Fcertifications\u002Fazure-ai-engineer)\n\n- [Azure机器学习训练与部署](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fdevops\u002Fpipelines\u002Ftargets\u002Fazure-machine-learning)\n\n- [从谷歌云培训学习机器学习和人工智能](https:\u002F\u002Fcloud.google.com\u002Ftraining\u002Fmachinelearning-ai)\n\n- [谷歌云机器学习速成课程](https:\u002F\u002Fdevelopers.google.com\u002Fmachine-learning\u002Fcrash-course\u002F)\n\n- [Udemy在线机器学习课程](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fmachine-learning\u002F)\n\n- [Coursera在线机器学习课程](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=machine%20learning&)\n\n- [edX在线课程与课堂学习机器学习](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fmachine-learning)\n\n### 图书\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n* [机器学习导论（PDF）](https:\u002F\u002Fai.stanford.edu\u002F~nilsson\u002FMLBOOK.pdf)\n\n* [《人工智能：一种现代方法》斯图尔特·J·拉塞尔和彼得·诺维格著](https:\u002F\u002Fwww.amazon.com\u002FArtificial-Intelligence-A-Modern-Approach\u002Fdp\u002F0134610997\u002Fref=sr_1_1?dchild=1&keywords=artificial+intelligence+a+modern+approach&qid=1626728093&sr=8-1)\n\n* [《深度学习》伊恩·古德费洛、约书亚·本吉奥和亚伦·库维尔著](https:\u002F\u002Fwww.deeplearningbook.org\u002F)\n\n* [《百页机器学习书》安德里·布尔科夫著](https:\u002F\u002Fthemlbook.com\u002Fwiki\u002Fdoku.php)\n\n    - [GitHub上的百页机器学习书](https:\u002F\u002Fgithub.com\u002Faburkov\u002FtheMLbook)\n\n* [汤姆·M·米切尔的《机器学习》](https:\u002F\u002Fwww.cs.cmu.edu\u002F~tom\u002FNewChapters.html)\n\n* [《编程集体智慧：构建智能Web 2.0应用》托比·塞加兰著](https:\u002F\u002Fwww.amazon.com\u002FProgramming-Collective-Intelligence-Building-Applications\u002Fdp\u002F0596529325\u002Fref=sr_1_1?crid=8EI42XMXESGB&keywords=Programming+Collective+Intelligence%3A+Building+Smart+Web+2.0+Applications&qid=1654318595&sprefix=programming+collective+intelligence+building+smart+web+2.0+applications%2Caps%2C194&sr=8-1)\n\n* [《机器学习：算法视角，第二版》](https:\u002F\u002Fwww.amazon.com\u002FMachine-Learning-Algorithmic-Perspective-Recognition\u002Fdp\u002F1466583282\u002Fref=sr_1_8?crid=2RIQ8OMMASS3&keywords=Pattern+Recognition+and+Machine+Learning&qid=1654318681&sprefix=pattern+recognition+and+machine+learning%2Caps%2C184&sr=8-8)\n\n* [克里斯托弗·M·毕晓普的《模式识别与机器学习》](https:\u002F\u002Fwww.amazon.com\u002FPattern-Recognition-Learning-Information-Statistics\u002Fdp\u002F1493938436\u002Fref=sr_1_4?crid=2RIQ8OMMASS3&keywords=Pattern+Recognition+and+Machine+Learning&qid=1654318681&sprefix=pattern+recognition+and+machine+learning%2Caps%2C184&sr=8-4)\n\n* [史蒂文·伯德、伊万·克莱因和爱德华·洛珀合著的《用Python进行自然语言处理》](https:\u002F\u002Fwww.amazon.com\u002FNatural-Language-Processing-Python-Analyzing\u002Fdp\u002F0596516495\u002Fref=sr_1_1?crid=O4XSCF3CNIBN&keywords=Natural+Language+Processing+with+Python&qid=1654318757&sprefix=natural+language+processing+with+python%2Caps%2C285&sr=8-1)\n\n* [莱昂纳德·埃迪森的《Python机器学习：面向初学者的技术指南》](https:\u002F\u002Fwww.amazon.com\u002FPython-Machine-Learning-Technical-Beginners\u002Fdp\u002F1986340872\u002Fref=sr_1_1?crid=1W5X2WV05GDQK&keywords=Python+Machine+Learning%3A+A+Technical+Approach+to+Machine+Learning+for+Beginners&qid=1654318782&sprefix=python+machine+learning+a+technical+approach+to+machine+learning+for+beginners%2Caps%2C212&sr=8-1)\n\n* [戴维·巴伯的《贝叶斯推理与机器学习》](https:\u002F\u002Fwww.amazon.com\u002FBayesian-Reasoning-Machine-Learning-Barber\u002Fdp\u002F0521518148\u002Fref=sr_1_1?crid=1J054T5MUCD20&keywords=Bayesian+Reasoning+and+Machine+Learning&qid=1654318807&sprefix=bayesian+reasoning+and+machine+learning%2Caps%2C179&sr=8-1)\n\n* [奥利弗·西奥巴尔德的《给绝对初学者的机器学习：纯英文入门》](https:\u002F\u002Fwww.amazon.com\u002FMachine-Learning-Absolute-Beginners-Introduction-ebook\u002Fdp\u002FB08RWBSKQB\u002Fref=sr_1_1?crid=1JBS4KEHTY6I5&keywords=Machine+Learning+for+Absolute+Beginners%3A+A+Plain+English+Introduction&qid=1654318861&sprefix=machine+learning+for+absolute+beginners+a+plain+english+introduction%2Caps%2C168&sr=8-1)\n\n* [本·威尔逊的《机器学习实战》](https:\u002F\u002Fwww.amazon.com\u002FMachine-Learning-Engineering-Action-Wilson\u002Fdp\u002F1617298719\u002Fref=sr_1_1?crid=6S9F2MJHAQX1&keywords=Machine+Learning+in+Action&qid=1654318897&sprefix=machine+learning+in+action%2Caps%2C174&sr=8-1)\n\n* [奥雷利安·热隆的《动手学机器学习：使用Scikit-Learn、Keras和TensorFlow构建智能系统》](https:\u002F\u002Fwww.amazon.com\u002FHands-Machine-Learning-Scikit-Learn-TensorFlow\u002Fdp\u002F1492032646\u002Fref=sr_1_6?crid=2RIQ8OMMASS3&keywords=Pattern+Recognition+and+Machine+Learning&qid=1654318681&sprefix=pattern+recognition+and+machine+learning%2Caps%2C184&sr=8-6)\n\n* [安德烈亚斯·C·穆勒和萨拉·圭多合著的《用Python入门机器学习：数据科学家指南》](https:\u002F\u002Fwww.amazon.com\u002FIntroduction-Machine-Learning-Python-Scientists\u002Fdp\u002F1449369413\u002Fref=sr_1_1?crid=3SGFHBBU06GB6&keywords=Introduction+to+Machine+Learning+with+Python%3A+A+Guide+for+Data+Scientists&qid=1654318969&sprefix=introduction+to+machine+learning+with+python+a+guide+for+data+scientists%2Caps%2C181&sr=8-1)\n\n* [德鲁·康威和约翰·迈尔斯·怀特合著的《黑客的机器学习：案例研究与算法，助你入门》](https:\u002F\u002Fwww.amazon.com\u002FMachine-Learning-Hackers-Studies-Algorithms\u002Fdp\u002F1449303714\u002Fref=sr_1_1?crid=2PQABQ4T9B8K5&keywords=Machine+Learning+for+Hackers%3A+Case+Studies+and+Algorithms+to+Get+you+Started&qid=1654318629&sprefix=machine+learning+for+hackers+case+studies+and+algorithms+to+get+you+started%2Caps%2C162&sr=8-1)\n\n* [特雷弗·哈斯蒂、罗伯特·蒂布希拉尼和杰罗姆·弗里德曼合著的《统计学习要素：数据挖掘、推断与预测》](https:\u002F\u002Fwww.amazon.com\u002FElements-Statistical-Learning-Prediction-Statistics\u002Fdp\u002F0387848576\u002Fref=sr_1_1?crid=1HOK9M9GFHTK9&keywords=The+Elements+of+Statistical+Learning%3A+Data+Mining%2C+Inference%2C+and+Prediction&qid=1654318661&sprefix=the+elements+of+statistical+learning+data+mining%2C+inference%2C+and+prediction+%2Caps%2C215&sr=8-1)\n\n* [分布式机器学习模式](https:\u002F\u002Fgithub.com\u002Fterrytangyuan\u002Fdistributed-ml-patterns) - 书籍（可在线免费阅读）+ 代码\n* [真实世界中的机器学习](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Freal-world-machine-learning) [免费章节]\n* [统计学习导论](https:\u002F\u002Fwww-bcf.usc.edu\u002F~gareth\u002FISL\u002F) - 书籍 + R 代码\n* [统计学习要素](https:\u002F\u002Fweb.stanford.edu\u002F~hastie\u002FElemStatLearn\u002F) - 书籍\n* [思考贝叶斯](https:\u002F\u002Fgreenteapress.com\u002Fwp\u002Fthink-bayes\u002F) - 书籍 + Python 代码\n* [挖掘海量数据集](https:\u002F\u002Finfolab.stanford.edu\u002F~ullman\u002Fmmds\u002Fbook.pdf)\n* [机器学习初探](https:\u002F\u002Fwww.ics.uci.edu\u002F~welling\u002Fteaching\u002F273ASpring10\u002FIntroMLBook.pdf)\n* [机器学习导论](https:\u002F\u002Falex.smola.org\u002Fdrafts\u002Fthebook.pdf) - 亚历克斯·斯莫拉和 S.V.N. 维什瓦纳坦\n* [模式识别的概率理论](https:\u002F\u002Fwww.szit.bme.hu\u002F~gyorfi\u002Fpbook.pdf)\n* [信息检索导论](https:\u002F\u002Fnlp.stanford.edu\u002FIR-book\u002Fpdf\u002Firbookprint.pdf)\n* [预测：原理与实践](https:\u002F\u002Fotexts.com\u002Ffpp2\u002F)\n* [机器学习导论](https:\u002F\u002Farxiv.org\u002Fpdf\u002F0904.3664v1.pdf) - 阿姆农·沙舒阿\n* [强化学习](https:\u002F\u002Fwww.intechopen.com\u002Fbooks\u002Freinforcement_learning)\n* [机器学习](https:\u002F\u002Fwww.intechopen.com\u002Fbooks\u002Fmachine_learning)\n* [人工智能探索](https:\u002F\u002Fai.stanford.edu\u002F~nilsson\u002FQAI\u002Fqai.pdf)\n* [R 语言在数据科学中的应用](https:\u002F\u002Fleanpub.com\u002Frprogramming)\n* [数据挖掘——实用的机器学习工具与技术](https:\u002F\u002Fcdn.preterhuman.net\u002Ftexts\u002Fscience_and_technology\u002Fartificial_intelligence\u002FData%20Mining%20Practical%20Machine%20Learning%20Tools%20and%20Techniques%202d%20ed%20-%20Morgan%20Kaufmann.pdf)\n* [使用 TensorFlow 的机器学习](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fmachine-learning-with-tensorflow)\n* [机器学习系统](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fmachine-learning-systems)\n* [机器学习基础](https:\u002F\u002Fcs.nyu.edu\u002F~mohri\u002Fmlbook\u002F) - 梅赫里亚尔·莫赫里、阿夫辛·罗斯塔米扎德和阿米特·塔尔瓦卡尔\n* [人工智能驱动的搜索](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fai-powered-search) - 特雷·格兰杰、道格·特恩布尔、麦克斯·欧文\n* [机器学习集成方法](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fensemble-methods-for-machine-learning) - 戈塔姆·库纳普利\n* [机器学习工程实战](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fmachine-learning-engineering-in-action) - 本·威尔逊\n* [隐私保护的机器学习](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fprivacy-preserving-machine-learning) - J. 莫里斯·张、迪·庄、G. 杜明杜·萨马拉维拉\n* [自动化机器学习实战](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fautomated-machine-learning-in-action) - 宋清泉、金海峰和胡霞\n* [分布式机器学习模式](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fdistributed-machine-learning-patterns) - 唐源\n* [管理机器学习项目：从设计到部署](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fmanaging-machine-learning-projects) - 西蒙·汤普森\n* [因果机器学习](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fcausal-machine-learning) - 罗伯特·内斯\n* [贝叶斯优化实战](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fbayesian-optimization-in-action) - 全阮\n* [深入机器学习算法](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fmachine-learning-algorithms-in-depth) - 瓦季姆·斯莫利亚科夫\n* [优化算法](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Foptimization-algorithms) - 阿拉·卡米斯\n* [实用梯度提升](https:\u002F\u002Fwww.amazon.com\u002Fdp\u002FB0BL1HRD6Z) 由纪尧姆·索潘著\n\n### YouTube 教程\n\n[返回顶部](#table-of-contents)\n\n[![吴恩达：人工智能的机遇 - 斯坦福大学 2023 年](https:\u002F\u002Fytcards.demolab.com\u002F?id=5p248yoa3oE&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"Andrew Ng: Opportunities in AI - Standford 2023\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=5p248yoa3oE)\n[![人工智能究竟是如何工作的？](https:\u002F\u002Fytcards.demolab.com\u002F?id=3ihjz7g1OQM&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"How Does AI Actually Work?\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=3ihjz7g1OQM)\n[![人工智能与机器学习](https:\u002F\u002Fytcards.demolab.com\u002F?id=4RixMPF4xis&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"AI vs Machine Learning\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=4RixMPF4xis)\n[![机器学习与深度学习](https:\u002F\u002Fytcards.demolab.com\u002F?id=q6kJ71tEYqM&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"Machine Learning vs Deep Learning\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=q6kJ71tEYqM)\n[![什么是 Transformer（机器学习模型）？](https:\u002F\u002Fytcards.demolab.com\u002F?id=ZXiruGOCn9s&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"What are Transformers (Machine Learning Model)?\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=ZXiruGOCn9s)\n[![那么，什么是神经网络呢？| 深度学习第一章](https:\u002F\u002Fytcards.demolab.com\u002F?id=aircAruvnKk&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"But what is a neural network? | Chapter 1, Deep learning\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=aircAruvnKk)\n[![给机器学习初学者的建议 | 安德烈·卡帕西和莱克斯·弗里德曼](https:\u002F\u002Fytcards.demolab.com\u002F?id=I2ZK3ngNvvI&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"Advice for machine learning beginners | Andrej Karpathy and Lex Fridman\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=I2ZK3ngNvvI)\n[![100 秒讲透机器学习](https:\u002F\u002Fytcards.demolab.com\u002F?id=PeMlggyqz0Y&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"Machine Learning Explained in 100 Seconds\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=PeMlggyqz0Y)\n[![2023 年如何学习人工智能和机器学习——完整路线图](https:\u002F\u002Fytcards.demolab.com\u002F?id=KEB-w9DUdCw&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"How to learn AI and ML in 2023 - A complete roadmap\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=KEB-w9DUdCw)\n[![PyTorch 深度学习与机器学习——完整课程](https:\u002F\u002Fytcards.demolab.com\u002F?id=V_xro1bcAuA&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"PyTorch for Deep Learning & Machine Learning – Full Course\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=V_xro1bcAuA)\n[![使用 Python 和 TensorFlow 进行计算机视觉深度学习](https:\u002F\u002Fytcards.demolab.com\u002F?id=IA3WxTTPXqQ&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"Deep Learning for Computer Vision with Python and TensorFlow\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=IA3WxTTPXqQ)\n[![大型语言模型的工作原理](https:\u002F\u002Fytcards.demolab.com\u002F?id=5sLYAQS9sWQ&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"How Large Language Models Work\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=5sLYAQS9sWQ)\n[![什么是大型语言模型（LLMs）？](https:\u002F\u002Fytcards.demolab.com\u002F?id=iR2O2GPbB0E&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"What are Large Language Models (LLMs)?\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=iR2O2GPbB0E)\n[![大型语言模型导论](https:\u002F\u002Fytcards.demolab.com\u002F?id=zizonToFXDs&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"Introduction to large language model\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=zizonToFXDs)\n[![用 Python 从零开始创建大型语言模型](https:\u002F\u002Fytcards.demolab.com\u002F?id=UU1WVnMk4E8&lang=en&background_color=%230d1117&title_color=%23ffffff&stats_color=%23dedede&width=240 \"Create a Large Language Model from Scratch with Python\")](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=UU1WVnMk4E8)\n\n# 机器学习框架、库和工具\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n[TensorFlow](https:\u002F\u002Fwww.tensorflow.org) 是一个端到端的开源机器学习平台。它拥有全面且灵活的工具、库和社区资源生态系统，使研究人员能够推动机器学习领域的前沿发展，也让开发者可以轻松构建和部署基于机器学习的应用程序。\n\n[Keras](https:\u002F\u002Fkeras.io) 是一个用 Python 编写的高级神经网络 API，可运行在 TensorFlow、CNTK 或 Theano 之上。它专注于支持快速实验，能够在 TensorFlow、Microsoft Cognitive Toolkit、R、Theano 或 PlaidML 等框架之上运行。\n\n[PyTorch](https:\u002F\u002Fpytorch.org) 是一个用于处理不规则输入数据（如图、点云和流形）的深度学习库。主要由 Facebook 的 AI 研究实验室开发。\n\n[Amazon SageMaker](https:\u002F\u002Faws.amazon.com\u002Fsagemaker\u002F) 是一项完全托管的服务，使每位开发者和数据科学家都能快速构建、训练和部署机器学习（ML）模型。SageMaker 可以减轻机器学习流程中每个步骤的繁重工作，从而更容易开发出高质量的模型。\n\n[Azure Databricks](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fdatabricks\u002F) 是一种基于 Apache Spark 的快速协作式大数据分析服务，专为数据科学和数据工程设计。Azure Databricks 可在几分钟内设置好 Apache Spark 环境，并实现自动扩展，在交互式工作区中进行共享项目协作。Azure Databricks 支持 Python、Scala、R、Java 和 SQL，以及 TensorFlow、PyTorch 和 scikit-learn 等数据科学框架和库。\n\n[Microsoft Cognitive Toolkit (CNTK)](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fcognitive-toolkit\u002F) 是一个面向商业级分布式深度学习的开源工具包。它通过有向图将神经网络描述为一系列计算步骤。CNTK 允许用户轻松实现并组合流行的模型类型，例如前馈 DNN、卷积神经网络（CNN）和循环神经网络（RNN\u002FLSTM）。CNTK 实现了随机梯度下降法（SGD，误差反向传播）的学习，并支持自动微分以及跨多个 GPU 和服务器的并行化。\n\n[Apple CoreML](https:\u002F\u002Fdeveloper.apple.com\u002Fdocumentation\u002Fcoreml) 是一个帮助将机器学习模型集成到应用中的框架。Core ML 为所有模型提供统一的表示形式。您的应用可以使用 Core ML API 和用户数据来进行预测，也可以在用户的设备上训练或微调模型。模型是通过将机器学习算法应用于一组训练数据而得到的结果。您可以使用模型根据新的输入数据做出预测。\n\n[Apache OpenNLP](https:\u002F\u002Fopennlp.apache.org\u002F) 是一个基于机器学习的自然语言文本处理开源工具库。它提供了一系列 API，可用于诸如 [命名实体识别](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FNamed-entity_recognition)、[句子检测]()、[词性标注](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FPart-of-speech_tagging)、[分词](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FTokenization_(data_security))、[特征提取](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FFeature_extraction)、[组块分析](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FChunking_(psychology))、[句法分析](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FParsing)以及 [共指消解](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FCoreference) 等任务。\n\n[Apache Airflow](https:\u002F\u002Fairflow.apache.org) 是一个由社区创建的开源工作流管理平台，用于以编程方式编写、调度和监控工作流。安装简便、原理清晰、可扩展性强。Airflow 具有模块化架构，并使用消息队列来协调任意数量的工作节点。Airflow 能够无限扩展。\n\n[开放神经网络交换格式（ONNX）](https:\u002F\u002Fgithub.com\u002Fonnx) 是一个开放生态系统，旨在帮助 AI 开发者在项目演进过程中选择合适的工具。ONNX 提供了一种针对深度学习和传统机器学习模型的开源格式，定义了一个可扩展的计算图模型，以及内置算子和标准数据类型的规范。\n\n[Apache MXNet](https:\u002F\u002Fmxnet.apache.org\u002F) 是一个专为高效性和灵活性设计的深度学习框架。它允许您混合使用符号式和命令式编程，从而最大化效率和生产力。MXNet 的核心是一个动态依赖调度器，能够实时自动并行化符号式和命令式操作。在其之上还有一层图优化层，使符号式执行既快速又节省内存。MXNet 具有良好的移植性和轻量级特性，能够有效扩展到多 GPU 和多台机器上。支持 Python、R、Julia、Scala、Go、JavaScript 等多种语言。\n\n[AutoGluon](https:\u002F\u002Fautogluon.mxnet.io\u002Findex.html) 是一个用于深度学习的工具包，能够自动化机器学习任务，让您轻松在应用中实现强大的预测性能。只需几行代码，您就可以在表格数据、图像和文本数据上训练并部署高精度的深度学习模型。\n\n[Anaconda](https:\u002F\u002Fwww.anaconda.com\u002F) 是一个非常流行的数据科学平台，适用于机器学习和深度学习，使用户能够开发、训练和部署模型。\n\n[PlaidML](https:\u002F\u002Fgithub.com\u002Fplaidml\u002Fplaidml) 是一款先进且可移植的张量编译器，能够在笔记本电脑、嵌入式设备或其他计算硬件支持不足或软件许可限制较多的设备上运行深度学习。\n\n[OpenCV](https:\u002F\u002Fopencv.org) 是一个高度优化的库，专注于实时计算机视觉应用。其 C++、Python 和 Java 接口支持 Linux、MacOS、Windows、iOS 和 Android 系统。\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) 是一个基于 SciPy、NumPy 和 matplotlib 构建的 Python 机器学习模块，使得许多流行的机器学习算法的稳健且简单的实现变得更加容易。\n\n[Weka](https:\u002F\u002Fwww.cs.waikato.ac.nz\u002Fml\u002Fweka\u002F) 是一款开源机器学习软件，可通过图形用户界面、标准终端应用程序或 Java API 访问。它广泛应用于教学、科研和工业领域，内置了大量用于标准机器学习任务的工具，并且还能透明地访问 scikit-learn、R 和 Deeplearning4j 等知名工具箱。\n\n[Caffe](https:\u002F\u002Fgithub.com\u002FBVLC\u002Fcaffe) 是一个以表达性、速度和模块化为核心理念的深度学习框架。它由伯克利人工智能研究实验室 (BAIR)、伯克利视觉与学习中心 (BVLC) 以及社区贡献者共同开发。\n\n[Theano](https:\u002F\u002Fgithub.com\u002FTheano\u002FTheano) 是一个 Python 库，允许您高效地定义、优化和评估涉及多维数组的数学表达式，并与 NumPy 紧密集成。\n\n[nGraph](https:\u002F\u002Fgithub.com\u002FNervanaSystems\u002Fngraph) 是一个用于深度学习的开源 C++ 库、编译器和运行时环境。nGraph 编译器旨在加速使用任何深度学习框架开发 AI 工作负载，并将其部署到各种硬件目标上。它为 AI 开发人员提供了自由度、高性能和易用性。\n\n[NVIDIA cuDNN](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcudnn) 是一个针对 [深度神经网络](https:\u002F\u002Fdeveloper.nvidia.com\u002Fdeep-learning) 的 GPU 加速原语库。cuDNN 提供了对前向和反向卷积、池化、归一化和激活层等标准操作的高度优化实现。cuDNN 可以加速广泛使用的深度学习框架，包括 [Caffe2](https:\u002F\u002Fcaffe2.ai\u002F)、[Chainer](https:\u002F\u002Fchainer.org\u002F)、[Keras](https:\u002F\u002Fkeras.io\u002F)、[MATLAB](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fdeep-learning.html)、[MxNet](https:\u002F\u002Fmxnet.incubator.apache.org\u002F)、[PyTorch](https:\u002F\u002Fpytorch.org\u002F) 和 [TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002F) 等。\n\n[Huginn](https:\u002F\u002Fgithub.com\u002Fhuginn\u002Fhuginn) 是一个自托管系统，用于构建代理程序，为您在线执行自动化任务。它可以抓取网页、监控事件，并代表您采取行动。Huginn 的代理会创建和消费事件，并沿着有向图传播这些事件。您可以把它看作是运行在您自己的服务器上的可 hack 版 IFTTT 或 Zapier。\n\n[Netron](https:\u002F\u002Fnetron.app\u002F) 是一款用于查看神经网络、深度学习和机器学习模型的工具。它支持 ONNX、TensorFlow Lite、Caffe、Keras、Darknet、PaddlePaddle、ncnn、MNN、Core ML、RKNN、MXNet、MindSpore Lite、TNN、Barracuda、Tengine、CNTK、TensorFlow.js、Caffe2 和 UFF 等格式。\n\n[Dopamine](https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fdopamine) 是一个用于快速原型化强化学习算法的研究框架。\n\n[DALI](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FDALI) 是一个 GPU 加速库，包含高度优化的数据处理构建模块和执行引擎，用于加速深度学习的训练和推理应用。\n\n[MindSpore Lite](https:\u002F\u002Fgithub.com\u002Fmindspore-ai\u002Fmindspore) 是一个新的开源深度学习训练\u002F推理框架，可用于移动、边缘和云场景。\n\n[Darknet](https:\u002F\u002Fgithub.com\u002Fpjreddie\u002Fdarknet) 是一个用 C 语言和 CUDA 编写的开源神经网络框架。它速度快、易于安装，并支持 CPU 和 GPU 计算。\n\n[PaddlePaddle](https:\u002F\u002Fgithub.com\u002FPaddlePaddle\u002FPaddle) 是一个易用、高效、灵活且可扩展的深度学习平台，最初由百度的科学家和工程师开发，旨在将深度学习应用于百度的众多产品中。\n\n[GoogleNotebookLM](https:\u002F\u002Fblog.google\u002Ftechnology\u002Fai\u002Fnotebooklm-google-ai\u002F) 是一款实验性的人工智能工具，利用语言模型的强大功能与您现有的内容相结合，以更快地获得关键见解。它类似于一位虚拟研究助理，可以根据您选择的资料来源总结事实、解释复杂概念，并头脑风暴新的关联。\n\n[Unilm](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Funilm) 是一个跨任务、跨语言和跨模态的大规模自监督预训练模型。\n\n[Semantic Kernel (SK)](https:\u002F\u002Faka.ms\u002Fsemantic-kernel) 是一个轻量级 SDK，能够将 AI 大型语言模型（LLMs）与传统编程语言集成。SK 的可扩展编程模型结合了自然语言语义函数、传统代码原生函数以及基于嵌入的记忆，从而释放新的潜力，为应用程序增添 AI 价值。\n\n[Pandas AI](https:\u002F\u002Fgithub.com\u002Fgventuri\u002Fpandas-ai) 是一个 Python 库，它将生成式人工智能能力集成到 Pandas 中，使数据框具备对话交互能力。\n\n[NCNN](https:\u002F\u002Fgithub.com\u002FTencent\u002Fncnn) 是一个针对移动平台优化的高性能神经网络推理框架。\n\n[MNN](https:\u002F\u002Fgithub.com\u002Falibaba\u002FMNN) 是一个极速、轻量级的深度学习框架，已在阿里巴巴的关键业务场景中经过严苛考验。\n\n[MediaPipe](https:\u002F\u002Fmediapipe.dev\u002F) 针对多种平台进行了端到端性能优化。查看演示 了解详情 将复杂的设备端机器学习简化 我们已抽象出使设备端机器学习可定制、生产就绪且跨平台可用的复杂性。\n\n[MegEngine](https:\u002F\u002Fgithub.com\u002FMegEngine) 是一个快速、可扩展且用户友好的深度学习框架，具有三大核心特性：统一的训练与推理框架。\n\n[ML.NET](https:\u002F\u002Fdot.net\u002Fml) 是一个机器学习库，被设计为一个可扩展的平台，允许您使用其他流行的机器学习框架（TensorFlow、ONNX、Infer.NET 等），并访问更多机器学习场景，如图像分类、目标检测等。\n\n[Ludwig](https:\u002F\u002Fludwig.ai\u002F) 是一个 [声明式机器学习框架](https:\u002F\u002Fludwig-ai.github.io\u002Fludwig-docs\u002Flatest\u002Fuser_guide\u002Fwhat_is_ludwig\u002F#why-declarative-machine-learning-systems)，它通过简单灵活的数据驱动配置系统，轻松定义机器学习工作流。\n\n[MMdnn](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002FMMdnn) 是一个全面且跨框架的工具，用于转换、可视化和诊断深度学习（DL）模型。“MM”代表模型管理，“dnn”是深度神经网络的缩写。可在 Caffe、Keras、MXNet、TensorFlow、CNTK、PyTorch、Onnx 和 CoreML 之间转换模型。\n\n[Horovod](https:\u002F\u002Fgithub.com\u002Fhorovod\u002Fhorovod) 是一个适用于 TensorFlow、Keras、PyTorch 和 Apache MXNet 的分布式深度学习训练框架。\n\n[Vaex](https:\u002F\u002Fvaex.io\u002F) 是一个高性能的 Python 库，用于处理惰性的大规模外存数据框（类似于 Pandas），以便可视化和探索大型表格数据集。\n\n[GluonTS](https:\u002F\u002Fts.gluon.ai\u002F) 是一个用于概率时间序列建模的 Python 包，专注于基于深度学习的模型，依托 [PyTorch](https:\u002F\u002Fpytorch.org\u002F) 和 [MXNet](https:\u002F\u002Fmxnet.apache.org\u002F) 构建。\n\n[MindsDB](http:\u002F\u002Fmindsdb.com\u002F) 是一个 ML-SQL 服务器，它允许使用 SQL 在最强大的数据库和数据仓库中运行机器学习工作流。\n\n[Jupyter Notebook](https:\u002F\u002Fjupyter.org\u002F) 是一个开源 Web 应用程序，允许您创建和共享包含实时代码、方程、可视化效果和叙述性文本的文档。Jupyter 广泛应用于数据清洗与转换、数值模拟、统计建模、数据可视化、数据科学和机器学习等领域。\n\n[Apache Spark](https:\u002F\u002Fspark.apache.org\u002F) 是一个用于大规模数据处理的统一分析引擎。它提供 Scala、Java、Python 和 R 等高级 API，以及一个优化的引擎，支持通用计算图进行数据分析。此外，它还支持丰富的高级工具，包括用于 SQL 和 DataFrame 的 Spark SQL、用于机器学习的 MLlib、用于图处理的 GraphX，以及用于流处理的 Structured Streaming。\n\n[Apache Spark Connector for SQL Server and Azure SQL](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fsql-spark-connector) 是一个高性能连接器，使您能够在大数据分析中使用事务性数据，并将结果持久化用于即席查询或报告。该连接器允许您将任何 SQL 数据库——无论是本地部署还是云端——用作 Spark 作业的输入数据源或输出数据接收端。\n\n[Apache PredictionIO](https:\u002F\u002Fpredictionio.apache.org\u002F) 是一个面向开发者、数据科学家和最终用户的开源机器学习框架。它支持事件收集、算法部署、评估以及通过 REST API 查询预测结果。该框架基于 Hadoop、HBase（以及其他数据库）、Elasticsearch、Spark 等可扩展的开源服务，并实现了所谓的 Lambda 架构。\n\n[Cluster Manager for Apache Kafka(CMAK)](https:\u002F\u002Fgithub.com\u002Fyahoo\u002FCMAK) 是一个用于管理 [Apache Kafka](https:\u002F\u002Fkafka.apache.org\u002F) 集群的工具。\n\n[BigDL](https:\u002F\u002Fbigdl-project.github.io\u002F) 是一个适用于 Apache Spark 的分布式深度学习库。借助 BigDL，用户可以将他们的深度学习应用编写为标准的 Spark 程序，这些程序可以直接在现有的 Spark 或 Hadoop 集群上运行。\n\n[Eclipse Deeplearning4J (DL4J)](https:\u002F\u002Fdeeplearning4j.konduit.ai\u002F) 是一组项目，旨在满足基于 JVM（Scala、Kotlin、Clojure 和 Groovy）的深度学习应用的所有需求。这意味着从原始数据开始，无论其来源和格式如何，都可以对其进行加载和预处理，进而构建和调优各种简单及复杂的深度学习网络。\n\n[Tensorman](https:\u002F\u002Fgithub.com\u002Fpop-os\u002Ftensorman) 是由 [System76]( https:\u002F\u002Fsystem76.com) 开发的一个用于轻松管理 Tensorflow 容器的实用工具。Tensorman 允许 Tensorflow 在一个与系统其他部分隔离的环境中运行。这种虚拟环境可以独立于基础系统运行，使您能够在任何支持 Docker 运行时的 Linux 发行版上使用任意版本的 Tensorflow。\n\n[Numba](https:\u002F\u002Fgithub.com\u002Fnumba\u002Fnumba) 是由 Anaconda, Inc. 赞助的一个开源、支持 NumPy 的 Python 优化编译器。它利用 LLVM 编译器项目将 Python 语法转换为机器码。Numba 可以编译大量以数值计算为主的 Python 代码，包括许多 NumPy 函数。此外，Numba 还支持循环的自动并行化、生成 GPU 加速代码，以及创建 ufunc 和 C 回调函数。\n\n[Chainer](https:\u002F\u002Fchainer.org\u002F) 是一个基于 Python 的深度学习框架，旨在提供灵活性。它提供了基于“定义即运行”方法（动态计算图）的自动微分 API，以及面向对象的高级 API 来构建和训练神经网络。Chainer 还通过 [CuPy](https:\u002F\u002Fgithub.com\u002Fcupy\u002Fcupy) 支持 CUDA\u002FcuDNN，从而实现高性能的训练和推理。\n\n[XGBoost](https:\u002F\u002Fxgboost.readthedocs.io\u002F) 是一个优化的分布式梯度提升库，设计目标是高效、灵活且可移植。它实现了梯度提升框架下的机器学习算法。XGBoost 提供了一种并行树提升技术（也称为 GBDT 或 GBM），能够快速而准确地解决许多数据科学问题。它支持在多台机器上进行分布式训练，包括 AWS、GCE、Azure 和 Yarn 集群。此外，它还可以与 Flink、Spark 等云数据流系统集成。\n\n[cuML](https:\u002F\u002Fgithub.com\u002Frapidsai\u002Fcuml) 是一套库，实现了机器学习算法和数学基础函数，其 API 与其他 RAPIDS 项目兼容。cuML 使数据科学家、研究人员和软件工程师能够在 GPU 上运行传统的表格型机器学习任务，而无需深入了解 CUDA 编程细节。在大多数情况下，cuML 的 Python API 与 scikit-learn 的 API 完全匹配。\n\n[Emu](https:\u002F\u002Fcalebwin.github.io\u002Femu) 是一个面向 Rust 的 GPGPU 库，专注于可移植性、模块化和性能。它是在 WebGPU 之上构建的类似 CUDA 的计算专用抽象层，提供了特定功能，使 WebGPU 的使用体验更接近 CUDA。\n\n[Scalene](https:\u002F\u002Fgithub.com\u002Fplasma-umass\u002Fscalene) 是一款高性能的 Python CPU、GPU 和内存分析工具，具备其他 Python 性能分析工具所不具备的功能。它的运行速度比许多其他分析工具快几个数量级，同时提供更为详尽的信息。\n\n[MLpack](https:\u002F\u002Fmlpack.org\u002F) 是一个快速、灵活的 C++ 机器学习库，用 C++ 编写，并基于 [Armadillo](https:\u002F\u002Farma.sourceforge.net\u002F) 线性代数库、[ensmallen](https:\u002F\u002Fensmallen.org\u002F) 数值优化库以及部分 [Boost](https:\u002F\u002Fboost.org\u002F) 库构建而成。\n\n[Netron](https:\u002F\u002Fnetron.app\u002F) 是一个用于查看神经网络、深度学习和机器学习模型的工具。它支持 ONNX、TensorFlow Lite、Caffe、Keras、Darknet、PaddlePaddle、ncnn、MNN、Core ML、RKNN、MXNet、MindSpore Lite、TNN、Barracuda、Tengine、CNTK、TensorFlow.js、Caffe2 和 UFF 格式。\n\n[Lightning](https:\u002F\u002Fgithub.com\u002FLightning-AI\u002Flightning) 是一个用于构建和训练 PyTorch 模型的工具，可通过 Lightning App 模板将模型与机器学习生命周期连接起来，而无需自行处理基础设施、成本管理、扩展等问题。\n\n[OpenNN](https:\u002F\u002Fwww.opennn.net\u002F) 是一个用于机器学习的开源神经网络库。它包含复杂的算法和实用工具，可用于处理多种人工智能解决方案。\n\n[H2O](https:\u002F\u002Fh2o.ai\u002F) 是一个 AI 云平台，能够解决复杂的业务问题，并加速新想法的发现，同时提供易于理解且值得信赖的结果。\n\n[Gensim](https:\u002F\u002Fgithub.com\u002FRaRe-Technologies\u002Fgensim) 是一个用于主题建模、文档索引和大规模语料相似度检索的 Python 库。其目标用户是自然语言处理（NLP）和信息检索（IR）领域的从业者。\n\n[llama.cpp](https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fllama.cpp) 是 Facebook 的 LLaMA 模型在 C\u002FC++ 中的移植版本。\n\n[hmmlearn](https:\u002F\u002Fgithub.com\u002Fhmmlearn\u002Fhmmlearn) 是一组用于无监督学习和推断 [隐马尔可夫模型](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FHidden_Markov_model) 的算法。\n\n[Nextjournal](https:\u002F\u002Fnextjournal.com\u002F) 是一个用于可重复研究的笔记本环境。它可以运行任何可以放入 Docker 容器的内容。通过多语言笔记本、自动版本控制和实时协作，您可以改进工作流程；借助按需资源调配（包括 GPU 支持），节省时间和成本。\n\n[IPython](https:\u002F\u002Fipython.org\u002F) 提供了一个丰富的交互式计算架构，包括：\n\n  - 一个功能强大的交互式 shell。\n  - 一个用于 [Jupyter](https:\u002F\u002Fjupyter.org\u002F) 的内核。\n  - 对交互式数据可视化和 [GUI 工具包](https:\u002F\u002Fipython.org\u002Fipython-doc\u002Fstable\u002Finteractive\u002Freference.html#gui-event-loop-support) 的支持。\n  - 灵活且可嵌入的解释器，可加载到您自己的项目中。\n  - 易于使用且高性能的 [并行计算](https:\u002F\u002Fipyparallel.readthedocs.io\u002Fen\u002Flatest\u002F) 工具。\n\n[Veles](https:\u002F\u002Fgithub.com\u002FSamsung\u002Fveles) 是三星目前正在开发的用于快速深度学习应用开发的分布式平台。\n\n[DyNet](https:\u002F\u002Fgithub.com\u002Fclab\u002Fdynet) 是由卡内基梅隆大学及其他机构开发的神经网络库。它用 C++ 编写（并提供 Python 绑定），设计目标是在 CPU 或 GPU 上高效运行，并能很好地处理结构会随每次训练实例变化的动态网络。这类网络在自然语言处理任务中尤为重要，DyNet 已被用于构建最先进的句法解析、机器翻译、形态学屈折等系统，以及其他众多应用领域。\n\n[Ray](https:\u002F\u002Fgithub.com\u002Fray-project\u002Fray) 是一个用于扩展 AI 和 Python 应用程序的统一框架。它由核心分布式运行时和用于加速机器学习工作负载的库套件（Ray AIR）组成。\n\n[whisper.cpp](https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fwhisper.cpp) 是 OpenAI 的 Whisper 自动语音识别（ASR）模型的高性能推理实现。\n\n[ChatGPT Plus](https:\u002F\u002Fopenai.com\u002Fblog\u002Fchatgpt-plus\u002F) 是 ChatGPT 的一项试点订阅计划（**每月 20 美元**），该对话式 AI 能够与您聊天、回答后续问题，并对不正确的假设提出质疑。\n\n[Auto-GPT](https:\u002F\u002Fgithub.com\u002FSignificant-Gravitas\u002FAuto-GPT) 是一种“AI 代理”，它可以根据自然语言给出的目标，将其分解为子任务，并通过互联网和其他工具在自动化循环中尝试实现该目标。它使用 OpenAI 的 GPT-4 或 GPT-3.5 API，是首批利用 GPT-4 执行自主任务的应用之一。\n\n[mckaywrigley 制作的 Chatbot UI](https:\u002F\u002Fgithub.com\u002Fmckaywrigley\u002Fchatbot-ui) 是一个基于 Next.js、TypeScript 和 Tailwind CSS 构建的高级聊天机器人工具包，专为 OpenAI 的聊天模型设计，建立在 Chatbot UI Lite 之上。此版本的 ChatBot UI 同时支持 GPT-3.5 和 GPT-4 模型。对话会话存储在您的浏览器本地，您还可以导出和导入会话，以防止数据丢失。请参阅 [演示](https:\u002F\u002Ftwitter.com\u002Fmckaywrigley\u002Fstatus\u002F1636103188733640704)。\n\n[mckaywrigley 制作的 Chatbot UI Lite](https:\u002F\u002Fgithub.com\u002Fmckaywrigley\u002Fchatbot-ui-lite) 是一个简单的聊天机器人入门工具包，适用于 OpenAI 的聊天模型，使用 Next.js、TypeScript 和 Tailwind CSS 构建。请参阅 [演示](https:\u002F\u002Ftwitter.com\u002Fmckaywrigley\u002Fstatus\u002F1636103188733640704)。\n\n[MiniGPT-4](https:\u002F\u002Fminigpt-4.github.io\u002F) 是一种利用先进大型语言模型增强视觉-语言理解能力的系统。\n\n[GPT4All](https:\u002F\u002Fgithub.com\u002Fnomic-ai\u002Fgpt4all) 是一个开源聊天机器人生态系统，基于大量干净的助手数据进行训练，这些数据包括代码、故事和对话，其基础是 [LLaMa](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fllama) 模型。\n\n[GPT4All UI](https:\u002F\u002Fgithub.com\u002Fnomic-ai\u002Fgpt4all-ui) 是一个 Flask Web 应用程序，提供用于与 GPT4All 聊天机器人交互的聊天界面。\n\n[Alpaca.cpp](https:\u002F\u002Fgithub.com\u002Fantimatter15\u002Falpaca.cpp) 是一款可在您设备上本地运行的快速类 ChatGPT 模型。它结合了 [LLaMA 基础模型](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fllama) 以及对 [斯坦福 Alpaca](https:\u002F\u002Fgithub.com\u002Ftatsu-lab\u002Fstanford_alpaca) 的开源复现——该模型是对基础模型进行微调，使其能够遵循指令（类似于用于训练 ChatGPT 的 [RLHF](https:\u002F\u002Fhuggingface.co\u002Fblog\u002Frlhf) 技术），并针对 [llama.cpp](https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fllama.cpp) 进行了一系列修改，以添加聊天界面。\n\n[llama.cpp](https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fllama.cpp) 是 Facebook 的 LLaMA 模型在 C\u002FC++ 中的移植版本。\n\n[OpenPlayground](https:\u002F\u002Fgithub.com\u002Fnat\u002Fopenplayground) 是一个可在您设备上本地运行类 ChatGPT 模型的实验平台。\n\n[Vicuna](https:\u002F\u002Fvicuna.lmsys.org\u002F) 是一个通过微调 LLaMA 模型训练而成的开源聊天机器人。据称其性能可达到 ChatGPT 的 90% 以上，而训练成本仅为 300 美元。\n\n[Yeagar ai](https:\u002F\u002Fgithub.com\u002Fyeagerai\u002Fyeagerai-agent) 是一个 Langchain 代理创建工具，旨在帮助您轻松构建、原型化和部署由 AI 驱动的代理。\n\n[Vicuna](https:\u002F\u002Fvicuna.lmsys.org\u002F) 是通过微调 LLaMA 基础模型创建的，使用的数据约有 7 万个由用户在 ShareGPT.com 上分享的对话，这些对话通过公开 API 收集而来。为确保数据质量，系统会将 HTML 格式转换回 Markdown，并过滤掉一些不适当或低质量的样本。\n\n[ShareGPT](https:\u002F\u002Fsharegpt.com\u002F) 是一个只需点击即可分享您最疯狂的 ChatGPT 对话的地方。截至目前，已共享 198,404 条对话。\n\n[FastChat](https:\u002F\u002Fgithub.com\u002Flm-sys\u002FFastChat) 是一个用于训练、服务和评估基于大型语言模型的聊天机器人的开放平台。\n\n[Haystack](https:\u002F\u002Fhaystack.deepset.ai\u002F) 是一个开源 NLP 框架，允许您使用 Transformer 模型和 LLM（如 GPT-4、ChatGPT 等）与您的数据进行交互。它提供了生产就绪的工具，可快速构建复杂的决策制定、问答、语义搜索、文本生成等应用。\n\n[StableLM（Stability AI 语言模型）](https:\u002F\u002Fgithub.com\u002FStability-AI\u002FStableLM) 是一系列语言模型，未来将持续更新新的检查点。\n\n[Databricks’ Dolly](https:\u002F\u002Fgithub.com\u002Fdatabrickslabs\u002Fdolly) 是一个遵循指令的大型语言模型，由 Databricks 的机器学习平台训练而成，且获准用于商业用途。\n\n[GPTCach](https:\u002F\u002Fgptcache.readthedocs.io\u002F) 是一个用于为 LLM 查询创建语义缓存的库。\n\n[AlaC](https:\u002F\u002Fgithub.com\u002Fgofireflyio\u002Faiac) 是一个人工智能基础设施即代码生成器。\n\n[Adrenaline](https:\u002F\u002Fuseadrenaline.com\u002F) 是一款让您与代码库对话的工具，它基于静态分析、向量搜索和大型语言模型技术。\n\n[OpenAssistant](https:\u002F\u002Fopen-assistant.io\u002F) 是一个基于聊天的助手，能够理解任务、与第三方系统交互，并动态获取信息来完成任务。\n\n[DoctorGPT](https:\u002F\u002Fgithub.com\u002Fingyamilmolinar\u002Fdoctorgpt) 是一个轻量级的自包含二进制文件，用于监控应用程序日志中的问题并进行诊断。\n\n[HttpGPT](https:\u002F\u002Fgithub.com\u002Flucoiso\u002FUEHttpGPT\u002Freleases) 是一款虚幻引擎 5 插件，可通过异步 REST 请求方便地与 OpenAI 的基于 GPT 的服务（ChatGPT 和 DALL-E）集成，使开发者能够轻松与这些服务通信。它还包含编辑器工具，可直接在引擎中集成 ChatGPT 和 DALL-E 图像生成功能。\n\n[PaLM 2](https:\u002F\u002Fai.google\u002Fdiscover\u002Fpalm2) 是新一代大型语言模型，继承了 Google 在机器学习和负责任的人工智能领域取得的突破性研究成果。它具备先进的推理能力，涵盖代码和数学、分类与问答、翻译与多语言能力，以及比我们之前最先进的 LLM 更出色的自然语言生成能力。\n\n[Med-PaLM](https:\u002F\u002Fsites.research.google\u002Fmed-palm\u002F) 是一款大型语言模型 (LLM)，旨在为医学问题提供高质量的答案。它充分利用了 Google 的大型语言模型能力，并通过精心策划的医学专家演示将其与医学领域紧密结合。\n\n[Sec-PaLM](https:\u002F\u002Fcloud.google.com\u002Fblog\u002Fproducts\u002Fidentity-security\u002Frsa-google-cloud-security-ai-workbench-generative-ai) 是一类大型语言模型 (LLMs)，能够加速负责组织安全的人员的工作效率。这些新模型不仅让人们以更自然、更具创意的方式理解和管理安全性。\n\n### 大语言模型训练框架\n\n[返回顶部](#table-of-contents)\n \n - [Alpa](https:\u002F\u002Falpa.ai\u002Findex.html) 是一个用于训练和部署大规模神经网络的系统。\n - [BayLing](https:\u002F\u002Fgithub.com\u002Fictnlp\u002FBayLing) - 一款英中双语大语言模型，具备先进的语言对齐能力，在英中文本生成、指令遵循及多轮对话交互方面表现出色。\n - [BLOOM](https:\u002F\u002Fhuggingface.co\u002Fbigscience\u002Fbloom) - BigScience 大规模开源多语言语言模型 [BLOOM-LoRA](https:\u002F\u002Fgithub.com\u002Flinhduongtuan\u002FBLOOM-LORA)\n - [Cerebras-GPT](https:\u002F\u002Fwww.cerebras.net\u002Fblog\u002Fcerebras-gpt-a-family-of-open-compute-efficient-large-language-models\u002F) - 一系列开放、计算高效的大型语言模型。\n - [DeepSpeed](https:\u002F\u002Fwww.deepspeed.ai\u002F) 是一个深度学习优化库，旨在使分布式训练和推理变得简单、高效且有效。\n - [FairScale](https:\u002F\u002Ffairscale.readthedocs.io\u002Fen\u002Flatest\u002Fwhat_is_fairscale.html) 是一个用于高性能和大规模训练的 PyTorch 扩展库。该库在扩展 PyTorch 基础功能的同时，引入了多项最先进的模型并行化技术。\n - [GLM](https:\u002F\u002Fgithub.com\u002FTHUDM\u002FGLM) - GLM 是一种基于自回归填空目标预训练的通用语言模型，可针对各类自然语言理解与生成任务进行微调。\n - [OpenFlamingo](https:\u002F\u002Fgithub.com\u002Fmlfoundations\u002Fopen_flamingo) 是 DeepMind 的 [Flamingo](https:\u002F\u002Fwww.deepmind.com\u002Fblog\u002Ftackling-multiple-tasks-with-a-single-visual-language-model) 模型的开源实现框架，用于训练大型多模态模型。\n - [OPT](https:\u002F\u002Farxiv.org\u002Fabs\u002F2205.01068) - 开放式预训练 Transformer 语言模型。\n - [StarCoder](https:\u002F\u002Fhuggingface.co\u002Fblog\u002Fstarcoder) - Hugging Face 的代码专用大语言模型。\n - [UltraLM](https:\u002F\u002Fgithub.com\u002Fthunlp\u002FUltraChat) - 大规模、信息丰富且多样化的多轮对话模型。\n - [UL2](https:\u002F\u002Farxiv.org\u002Fabs\u002F2205.05131v1) - 一种统一的预训练框架，适用于跨数据集和不同设置的通用模型。\n\n### 大语言模型部署工具\n\n[返回顶部](#table-of-contents)\n\n- [Agenta](https:\u002F\u002Fgithub.com\u002Fagenta-ai\u002Fagenta) - 轻松构建、版本管理、评估和部署基于大语言模型的应用程序。\n- [BentoML](https:\u002F\u002Fbentoml.com\u002F) 专为基于大语言模型的应用而设计。\n- [CometLLM](https:\u002F\u002Fgithub.com\u002Fcomet-ml\u002Fcomet-llm) - 一个开源的 LLMOps 平台，用于记录、管理和可视化您的大语言模型提示及链路。跟踪提示模板、提示变量、提示执行时间、令牌使用情况等元数据，并可在单一界面中对提示输出进行评分及聊天历史可视化。\n- [FastChat](https:\u002F\u002Fgithub.com\u002Flm-sys\u002FFastChat) - 一个具有 Web UI 和 OpenAI 兼容 RESTful API 的分布式多模型大语言模型服务系统。\n- [Embedchain](https:\u002F\u002Fgithub.com\u002Fembedchain\u002Fembedchain) - 一个框架，用于基于您的数据集创建类似 ChatGPT 的聊天机器人。\n- [IntelliServer](https:\u002F\u002Fgithub.com\u002Fintelligentnode\u002FIntelliServer) - 通过提供统一的微服务来访问和测试多种 AI 模型，简化大语言模型的评估流程。\n- [Haystack](https:\u002F\u002Fhaystack.deepset.ai\u002F) - 一个开源的 NLP 框架，允许您使用来自 Hugging Face、OpenAI 和 Cohere 的大语言模型及 Transformer 模型与您自己的数据进行交互。\n- [Langroid](https:\u002F\u002Fgithub.com\u002Flangroid\u002Flangroid) - 通过多智能体编程充分利用大语言模型。\n- [LangChain](https:\u002F\u002Fgithub.com\u002Fhwchase17\u002Flangchain) - 通过组合性构建基于大语言模型的应用。\n- [LiteChain](https:\u002F\u002Fgithub.com\u002Frogeriochaves\u002Flitechain) - LangChain 的轻量级替代方案，用于组合大语言模型。\n- [Magentic](https:\u002F\u002Fgithub.com\u002Fjackmpcollins\u002Fmagentic) - 可以将大语言模型无缝集成为 Python 函数。\n- [Promptfoo](https:\u002F\u002Fgithub.com\u002Ftyppo\u002Fpromptfoo) - 测试您的提示词，评估和比较大语言模型的输出，捕捉回归问题并提升提示质量。\n- [OpenLLM](https:\u002F\u002Fgithub.com\u002Fbentoml\u002FOpenLLM) 是一个用于在生产环境中运行大型语言模型（LLMs）的开放平台。您可以轻松地对任何 LLM 进行微调、服务、部署和监控。\n- [Serge](https:\u002F\u002Fgithub.com\u002Fserge-chat\u002Fserge) - 一个基于 llama.cpp 构建的聊天界面，用于运行 Alpaca 模型。无需 API 密钥，完全自托管！\n- [SkyPilot](https:\u002F\u002Fgithub.com\u002Fskypilot-org\u002Fskypilot) - 在任意云平台上运行大语言模型和批处理任务。通过简单的界面获得最大成本节约、最高 GPU 可用性以及托管式执行。\n- [Text Generation Inference](https:\u002F\u002Fgithub.com\u002Fhuggingface\u002Ftext-generation-inference) - 一个基于 Rust、Python 和 gRPC 的文本生成推理服务器。该服务器在 [HuggingFace](https:\u002F\u002Fhuggingface.co\u002F) 的生产环境中用于支持 LLM 的 API 推理组件。\n- [vLLM](https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm) - 一个高吞吐量、内存高效的 LLM 推理与服务引擎。\n\n### 在本地运行大语言模型\n\n[返回顶部](#table-of-contents)\n\n * [在本地运行 Llama 2 的全面指南](https:\u002F\u002Freplicate.com\u002Fblog\u002Frun-llama-locally)\n * [lmsys.org 的排行榜](https:\u002F\u002Fchat.lmsys.org\u002F?leaderboard)\n * [LLM-Leaderboard](https:\u002F\u002Fgithub.com\u002FLudwigStumpp\u002Fllm-leaderboard)\n * [Hugging Face 的 Open LLM 排行榜](https:\u002F\u002Fhuggingface.co\u002Fspaces\u002FHuggingFaceH4\u002Fopen_llm_leaderboard)\n * [语言模型综合评估 (HELM)](https:\u002F\u002Fcrfm.stanford.edu\u002Fhelm\u002Flatest\u002F?groups=1)\n * [TextSynth Server 基准测试](https:\u002F\u002Fbellard.org\u002Fts_server\u002F)\n\n[LocalAI](https:\u002F\u002Flocalai.io\u002F) 是一个由社区驱动的自托管本地 OpenAI 兼容 API。它可作为 OpenAI 的直接替代品，在消费级硬件上运行大语言模型，无需 GPU。该 API 支持运行 ggml 兼容模型，包括 llama、gpt4all、rwkv、whisper、vicuna、koala、gpt4all-j、cerebras、falcon、dolly、starcoder 等多种模型。\n\n[llama.cpp](https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fllama.cpp) 是 Facebook 的 LLaMA 模型在 C\u002FC++ 中的移植版本。\n\n[ollama](https:\u002F\u002Follama.ai\u002F) 是一款工具，可以帮助用户在本地快速启动并运行 Llama 2 及其他大型语言模型。\n\n[LocalAI](https:\u002F\u002Flocalai.io\u002F) 是一个由社区驱动的自托管本地 OpenAI 兼容 API。它可作为 OpenAI 的直接替代品，在消费级硬件上运行大语言模型，无需 GPU。该 API 支持运行 ggml 兼容模型，包括 llama、gpt4all、rwkv、whisper、vicuna、koala、gpt4all-j、cerebras、falcon、dolly、starcoder 等多种模型。\n \n[Serge](https:\u002F\u002Fgithub.com\u002Fserge-chat\u002Fserge) 是一个通过 llama.cpp 与 Alpaca 对话的 Web 界面。完全自托管且容器化，提供易于使用的 API。 \n\n[OpenLLM](https:\u002F\u002Fgithub.com\u002Fbentoml\u002FOpenLLM) 是一个用于在生产环境中运行大型语言模型（LLMs）的开放平台。可以轻松地对任何 LLM 进行微调、服务、部署和监控。\n\n[Llama-gpt](https:\u002F\u002Fgithub.com\u002Fgetumbrel\u002Fllama-gpt) 是一款自托管的离线类 ChatGPT 聊天机器人。基于 Llama 2 构建，100% 私密，数据不会离开您的设备。 \n\n[Llama2 webui](https:\u002F\u002Fgithub.com\u002Fliltom-eth\u002Fllama2-webui) 是一款工具，可在任何地方（Linux\u002FWindows\u002FMac）使用 Gradio 界面，在 GPU 或 CPU 上本地运行任意 Llama 2 模型。您可以将 `llama2-wrapper` 用作生成式代理或应用的本地 Llama 2 后端。\n\n[Llama2.c](https:\u002F\u002Fgithub.com\u002Fkarpathy\u002Fllama2.c) 是一种工具，可以在 PyTorch 中训练 Llama 2 的 LLM 架构，然后使用一个简单的 700 行 C 文件（[run.c](https:\u002F\u002Fgithub.com\u002Fkarpathy\u002Fllama2.c\u002Fblob\u002Fmaster\u002Frun.c)）进行推理。\n\n[Alpaca.cpp](https:\u002F\u002Fgithub.com\u002Fantimatter15\u002Falpaca.cpp) 是一款在您设备上本地运行的快速类 ChatGPT 模型。它结合了 [LLaMA 基础模型](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fllama) 和 [斯坦福 Alpaca](https:\u002F\u002Fgithub.com\u002Ftatsu-lab\u002Fstanford_alpaca) 的开源复现版（[tloen\u002Falpaca-lora](https:\u002F\u002Fgithub.com\u002Ftloen\u002Falpaca-lora)），后者是对基础模型进行指令遵循的微调（类似于用于训练 ChatGPT 的 [RLHF](https:\u002F\u002Fhuggingface.co\u002Fblog\u002Frlhf)），并针对 [llama.cpp](https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fllama.cpp) 进行了一系列修改，以添加聊天界面。\n\n[GPT4All](https:\u002F\u002Fgithub.com\u002Fnomic-ai\u002Fgpt4all) 是一个开源聊天机器人生态系统，基于大量干净的助手数据集进行训练，这些数据集包括代码、故事和对话，其基础是 [LLaMa](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fllama)。\n\n[MiniGPT-4](https:\u002F\u002Fminigpt-4.github.io\u002F) 是一种利用先进大型语言模型增强视觉-语言理解能力的工具。\n\n[LoLLMS WebUI](https:\u002F\u002Fgithub.com\u002FParisNeo\u002Flollms-webui) 是一个大型语言模型（LLM）的枢纽。它的目标是提供一个用户友好的界面，以便访问和使用各种 LLM 模型来完成广泛的任务。无论您需要写作、编程、整理数据、生成图像，还是寻求问题的答案，都可以在这里找到帮助。\n\n[LM Studio](https:\u002F\u002Flmstudio.ai\u002F) 是一款用于发现、下载和本地运行 LLM 的工具。\n\n[Gradio Web UI](https:\u002F\u002Fgithub.com\u002Foobabooga\u002Ftext-generation-webui) 是一款面向大型语言模型的工具。支持 transformers、GPTQ、llama.cpp（ggml\u002Fgguf）以及 Llama 模型。\n\n[OpenPlayground](https:\u002F\u002Fgithub.com\u002Fnat\u002Fopenplayground) 是一个在您设备上本地运行类 ChatGPT 模型的实验平台。\n\n[Vicuna](https:\u002F\u002Fvicuna.lmsys.org\u002F) 是一个通过微调 LLaMA 训练而成的开源聊天机器人。据称其质量超过 ChatGPT 的 90%，而训练成本仅为 300 美元。\n\n[Yeagar ai](https:\u002F\u002Fgithub.com\u002Fyeagerai\u002Fyeagerai-agent) 是一个 Langchain 代理创建工具，旨在帮助您轻松构建、原型设计和部署人工智能驱动的代理。\n\n[KoboldCpp](https:\u002F\u002Fgithub.com\u002FLostRuins\u002Fkoboldcpp) 是一款易于使用的 GGML 模型 AI 文本生成软件。它是由 Concedo 发布的一个独立可分发包，基于 llama.cpp 构建，并添加了一个多功能的 Kobold API 端点、额外的格式支持、向后兼容性，以及一个功能丰富的用户界面，包含持久化的对话记录、编辑工具、保存格式、记忆、世界信息、作者注释、角色和场景等功能。\n\n# 算法\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n【模糊逻辑】是一种启发式方法，它能够实现更高级的决策树处理，并更好地与基于规则的编程相结合。\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_751e3caad00a.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**模糊逻辑系统的架构。来源：[ResearchGate](https:\u002F\u002Fwww.researchgate.net\u002Ffigure\u002FArchitecture-of-a-fuzzy-logic-system_fig2_309452475)**\n\n【支持向量机（SVM）】是一种监督学习模型，使用分类算法解决二分类问题。\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_db9febb4b018.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**支持向量机（SVM）。来源：[OpenClipArt](https:\u002F\u002Fopenclipart.org\u002Fdetail\u002F182977\u002Fsvm-support-vector-machines)**\n\n【神经网络】是机器学习的一个子集，也是深度学习算法的核心。其名称和结构灵感来源于人脑，模仿了生物神经元之间相互传递信号的过程。\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_4927bf734ef7.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**深度神经网络。来源：[IBM](https:\u002F\u002Fwww.ibm.com\u002Fcloud\u002Flearn\u002Fneural-networks)**\n\n【卷积神经网络（R-CNN）】是一种目标检测算法，它首先对图像进行分割以找到潜在的相关边界框，然后运行检测算法来识别这些边界框中最可能包含的对象。\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_8f5b375d6e6c.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**卷积神经网络。来源：[CS231n](https:\u002F\u002Fcs231n.github.io\u002Fconvolutional-networks\u002F#conv)**\n\n【循环神经网络（RNN）】是一类人工神经网络，适用于处理序列数据或时间序列数据。\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_ddd80740aee5.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**循环神经网络。来源：[Slideteam](https:\u002F\u002Fwww.slideteam.net\u002Frecurrent-neural-networks-rnns-ppt-powerpoint-presentation-file-templates.html)**\n\n【多层感知器（MLP）】是由多层【感知器】组成的多层神经网络，每层都带有阈值激活函数。\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_837d266dc173.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**多层感知器。来源：[DeepAI](https:\u002F\u002Fdeepai.org\u002Fmachine-learning-glossary-and-terms\u002Fmultilayer-perceptron)**\n\n【随机森林】是一种常用的机器学习算法，它通过结合多个决策树的输出来得出最终结果。森林中的每一棵决策树都不会被修剪，因此可以用于抽样和预测选择。由于其易用性和灵活性，随机森林在分类和回归问题中都得到了广泛应用。\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_5c55812f2c05.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**随机森林。来源：[wikimedia](https:\u002F\u002Fcommunity.tibco.com\u002Fwiki\u002Frandom-forest-template-tibco-spotfirer-wiki-page)**\n\n【决策树】是用于分类和回归的树状结构模型。\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_ee5160d69cb4.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**决策树。来源：[CMU](http:\u002F\u002Fwww.cs.cmu.edu\u002F~bhiksha\u002Fcourses\u002F10-601\u002Fdecisiontrees\u002F)**\n\n【朴素贝叶斯】是一种用于解决分类问题的机器学习算法。它基于【贝叶斯定理】，并假设各个特征之间具有较强的独立性。\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_50971225cc23.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**贝叶斯定理。来源：[mathisfun](https:\u002F\u002Fwww.mathsisfun.com\u002Fdata\u002Fbayes-theorem.html)**\n\n# PyTorch 开发\n\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_b48222faa05e.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## PyTorch 学习资源\n\n【PyTorch】是一个开源的深度学习框架，能够加速从研究到生产的流程，广泛应用于计算机视觉和自然语言处理等领域。PyTorch由【Facebook AI Research】实验室开发。\n\n【PyTorch 入门指南】(https:\u002F\u002Fpytorch.org\u002Fget-started\u002Flocally\u002F)\n\n【PyTorch 官方文档】(https:\u002F\u002Fpytorch.org\u002Fdocs\u002Fstable\u002Findex.html)\n\n【PyTorch 讨论论坛】(https:\u002F\u002Fdiscuss.pytorch.org\u002F)\n\n【Coursera 上的顶级 PyTorch 课程 | Coursera】(https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=pytorch&page=1)\n\n【Udemy 上的顶级 PyTorch 课程 | Udemy】(https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002FPyTorch\u002F)\n\n【通过在线课程和课堂学习 PyTorch | edX】(https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fpytorch)\n\n【PyTorch 基础知识 - Learn | Microsoft Docs】(https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Flearn\u002Fpaths\u002Fpytorch-fundamentals\u002F)\n\n【使用 PyTorch 进行深度学习入门 | Udacity】(https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fdeep-learning-pytorch--ud188)\n\n【在 Visual Studio Code 中进行 PyTorch 开发】(https:\u002F\u002Fcode.visualstudio.com\u002Fdocs\u002Fdatascience\u002Fpytorch-support)\n\n【Azure 上的 PyTorch - 使用 PyTorch 进行深度学习 | Microsoft Azure】(https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fdevelop\u002Fpytorch\u002F)\n\n【PyTorch - Azure Databricks | Microsoft Docs】(https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fdatabricks\u002Fapplications\u002Fmachine-learning\u002Ftrain-model\u002Fpytorch)\n\n【使用 PyTorch 进行深度学习 | Amazon Web Services (AWS)】(https:\u002F\u002Faws.amazon.com\u002Fpytorch\u002F)\n\n【在 Google Cloud 上开始使用 PyTorch】(https:\u002F\u002Fcloud.google.com\u002Fai-platform\u002Ftraining\u002Fdocs\u002Fgetting-started-pytorch)\n\n## PyTorch 工具、库和框架\n\n【PyTorch Mobile】(https:\u002F\u002Fpytorch.org\u002Fmobile\u002Fhome\u002F) 是一个端到端的机器学习工作流，涵盖从训练到部署的全过程，适用于 iOS 和 Android 移动设备。\n\n[TorchScript](https:\u002F\u002Fpytorch.org\u002Fdocs\u002Fstable\u002Fjit.html) 是一种从 PyTorch 代码中创建可序列化且可优化模型的方法。这使得任何 TorchScript 程序都可以从 Python 进程中保存，并在没有 Python 依赖的进程中加载。\n\n[TorchServe](https:\u002F\u002Fpytorch.org\u002Fserve\u002F) 是一个灵活且易于使用的工具，用于部署 PyTorch 模型。\n\n[Keras](https:\u002F\u002Fkeras.io) 是一个用 Python 编写的高级神经网络 API，可在 TensorFlow、CNTK 或 Theano 等后端上运行。它专注于支持快速实验，能够运行在 TensorFlow、Microsoft Cognitive Toolkit、R、Theano 或 PlaidML 等框架之上。\n\n[ONNX Runtime](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fonnxruntime) 是一个跨平台、高性能的机器学习推理和训练加速器。它支持来自 PyTorch 和 TensorFlow\u002FKeras 等深度学习框架的模型，以及 scikit-learn、LightGBM、XGBoost 等传统机器学习库中的模型。\n\n[Kornia](https:\u002F\u002Fkornia.github.io\u002F) 是一个可微分计算机视觉库，包含一系列用于解决通用计算机视觉问题的例程和可微模块。\n\n[PyTorch-NLP](https:\u002F\u002Fpytorchnlp.readthedocs.io\u002Fen\u002Flatest\u002F) 是一个用于 Python 中自然语言处理（NLP）的库。它基于最新的研究成果构建，自始至终都旨在支持快速原型开发。PyTorch-NLP 提供预训练的词嵌入、采样器、数据集加载器、评估指标、神经网络模块和文本编码器。\n\n[Ignite](https:\u002F\u002Fpytorch.org\u002Fignite) 是一个高级库，用于在 PyTorch 中灵活透明地训练和评估神经网络。\n\n[Hummingbird](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fhummingbird) 是一个将训练好的传统机器学习模型编译为张量计算的库。它允许用户无缝利用神经网络框架（如 PyTorch）来加速传统机器学习模型。\n\n[Deep Graph Library (DGL)](https:\u002F\u002Fwww.dgl.ai\u002F) 是一个基于 Python 的软件包，专为在 PyTorch 及其他框架之上轻松实现图神经网络模型系列而设计。\n\n[TensorLy](http:\u002F\u002Ftensorly.org\u002Fstable\u002Fhome.html) 是一个用于 Python 中张量方法和深度张量化神经网络的高级 API，旨在使张量学习变得简单。\n\n[GPyTorch](https:\u002F\u002Fcornellius-gp.github.io\u002F) 是一个使用 PyTorch 实现的高斯过程库，专为构建可扩展、灵活的高斯过程模型而设计。\n\n[Poutyne](https:\u002F\u002Fpoutyne.org\u002F) 是一个类似于 Keras 的 PyTorch 框架，可以处理训练神经网络所需的大量样板代码。\n\n[Forte](https:\u002F\u002Fgithub.com\u002Fasyml\u002Fforte\u002Ftree\u002Fmaster\u002Fdocs) 是一个用于构建 NLP 流水线的工具包，具有可组合的组件、便捷的数据接口以及跨任务交互功能。\n\n[TorchMetrics](https:\u002F\u002Fgithub.com\u002FPyTorchLightning\u002Fmetrics) 是一套用于分布式、可扩展 PyTorch 应用程序的机器学习指标。\n\n[Captum](https:\u002F\u002Fcaptum.ai\u002F) 是一个基于 PyTorch 构建的开源、可扩展的模型可解释性库。\n\n[Transformer](https:\u002F\u002Fgithub.com\u002Fhuggingface\u002Ftransformers) 是一个面向 PyTorch、TensorFlow 和 JAX 的最先进自然语言处理框架。\n\n[Hydra](https:\u002F\u002Fhydra.cc) 是一个用于优雅配置复杂应用的框架。\n\n[Accelerate](https:\u002F\u002Fhuggingface.co\u002Fdocs\u002Faccelerate) 是一种简单的方式，用于在多 GPU、TPU 上以混合精度训练和使用 PyTorch 模型。\n\n[Ray](https:\u002F\u002Fgithub.com\u002Fray-project\u002Fray) 是一个快速且简单的框架，用于构建和运行分布式应用程序。\n\n[ParlAI](http:\u002F\u002Fparl.ai\u002F) 是一个统一的平台，用于共享、训练和评估跨多种任务的对话模型。\n\n[PyTorchVideo](https:\u002F\u002Fpytorchvideo.org\u002F) 是一个用于视频理解研究的深度学习库。它提供了各种以视频为中心的模型、数据集、训练流水线等。\n\n[Opacus](https:\u002F\u002Fopacus.ai\u002F) 是一个允许使用差分隐私训练 PyTorch 模型的库。\n\n[PyTorch Lightning](https:\u002F\u002Fgithub.com\u002FwilliamFalcon\u002Fpytorch-lightning) 是一个类似于 Keras 的 PyTorch 机器学习库。它将核心的训练和验证逻辑交给你，而自动完成其余部分。\n\n[PyTorch Geometric Temporal](https:\u002F\u002Fgithub.com\u002Fbenedekrozemberczki\u002Fpytorch_geometric_temporal) 是 PyTorch Geometric 的时间（动态）扩展库。\n\n[PyTorch Geometric](https:\u002F\u002Fgithub.com\u002Frusty1s\u002Fpytorch_geometric) 是一个用于处理不规则输入数据（如图、点云和流形）的深度学习库。\n\n[Raster Vision](https:\u002F\u002Fdocs.rastervision.io\u002F) 是一个用于卫星和航空影像深度学习的开源框架。\n\n[CrypTen](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002FCrypTen) 是一个保护隐私的机器学习框架。其目标是使安全计算技术对机器学习从业者更加易用。\n\n[Optuna](https:\u002F\u002Foptuna.org\u002F) 是一个开源的超参数优化框架，用于自动化超参数搜索。\n\n[Pyro](http:\u002F\u002Fpyro.ai\u002F) 是一种用 Python 编写的通用概率编程语言（PPL），后端由 PyTorch 支持。\n\n[Albumentations](https:\u002F\u002Fgithub.com\u002Falbu\u002Falbumentations) 是一个快速且可扩展的图像增强库，适用于分类、分割、目标检测和姿态估计等不同的计算机视觉任务。\n\n[Skorch](https:\u002F\u002Fgithub.com\u002Fskorch-dev\u002Fskorch) 是一个与 scikit-learn  volfully 兼容的 PyTorch 高级库。\n\n[MMF](https:\u002F\u002Fmmf.sh\u002F) 是 Facebook AI Research (FAIR) 开发的一个用于视觉与语言多模态研究的模块化框架。\n\n[AdaptDL](https:\u002F\u002Fgithub.com\u002Fpetuum\u002Fadaptdl) 是一个资源自适应的深度学习训练和调度框架。\n\n[Polyaxon](https:\u002F\u002Fgithub.com\u002Fpolyaxon\u002Fpolyaxon) 是一个用于构建、训练和监控大规模深度学习应用的平台。\n\n[TextBrewer](http:\u002F\u002Ftextbrewer.hfl-rc.com\u002F) 是一个基于 PyTorch 的知识蒸馏工具包，用于自然语言处理。\n\n[AdverTorch](https:\u002F\u002Fgithub.com\u002FBorealisAI\u002Fadvertorch) 是一个用于对抗鲁棒性研究的工具箱。它包含生成对抗样本和防御攻击的模块。\n\n[NeMo](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNeMo) 是一个用于对话式 AI 的工具包。\n\n[ClinicaDL](https:\u002F\u002Fclinicadl.readthedocs.io\u002F) 是一个用于阿尔茨海默病可重复性分类的框架。\n\n[Stable Baselines3 (SB3)](https:\u002F\u002Fgithub.com\u002FDLR-RM\u002Fstable-baselines3) 是一组在 PyTorch 中可靠实现的强化学习算法。\n\n[TorchIO](https:\u002F\u002Fgithub.com\u002Ffepegar\u002Ftorchio) 是一套工具，用于高效地读取、预处理、采样、增强和写入深度学习应用中的 3D 医学图像。\n\n[PySyft](https:\u002F\u002Fgithub.com\u002FOpenMined\u002FPySyft) 是一个用于加密、保护隐私的深度学习的 Python 库。\n\n[Flair](https:\u002F\u002Fgithub.com\u002FflairNLP\u002Fflair) 是一个非常简单的框架，用于最先进的自然语言处理（NLP）。\n\n[Glow](https:\u002F\u002Fgithub.com\u002Fpytorch\u002Fglow) 是一个机器学习编译器，能够加速深度学习框架在不同硬件平台上的性能。\n\n[FairScale](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Ffairscale) 是一个 PyTorch 扩展库，用于在单台或多台机器\u002F节点上进行高性能和大规模训练。\n\n[MONAI](https:\u002F\u002Fmonai.io\u002F) 是一个深度学习框架，为医疗影像训练工作流的开发提供了领域优化的基础能力。\n\n[PFRL](https:\u002F\u002Fgithub.com\u002Fpfnet\u002Fpfrl) 是一个深度强化学习库，使用 PyTorch 以 Python 实现了多种最先进的深度强化学习算法。\n\n[Einops](https:\u002F\u002Fgithub.com\u002Farogozhnikov\u002Feinops) 是一种灵活而强大的张量操作库，旨在编写可读且可靠的代码。\n\n[PyTorch3D](https:\u002F\u002Fpytorch3d.org\u002F) 是一个深度学习库，为使用 PyTorch 进行 3D 计算机视觉研究提供了高效、可重用的组件。\n\n[Ensemble Pytorch](https:\u002F\u002Fensemble-pytorch.readthedocs.io\u002F) 是一个面向 PyTorch 的统一集成框架，用于提升深度学习模型的性能和鲁棒性。\n\n[Lightly](https:\u002F\u002Fgithub.com\u002Flightly-ai\u002Flightly) 是一个用于自监督学习的计算机视觉框架。\n\n[Higher](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fhigher) 是一个库，它使得实现任意复杂的基于梯度的元学习算法以及嵌套优化循环变得容易，同时几乎不改变原生 PyTorch 的代码结构。\n\n[Horovod](http:\u002F\u002Fhorovod.ai\u002F) 是一个针对深度学习框架的分布式训练库。Horovod 的目标是让分布式深度学习既快速又易于使用。\n\n[PennyLane](https:\u002F\u002Fpennylane.ai\u002F) 是一个用于量子机器学习、自动微分以及混合量子-经典计算优化的库。\n\n[Detectron2](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fdetectron2) 是 FAIR 推出的新一代目标检测与分割平台。\n\n[Fastai](https:\u002F\u002Fdocs.fast.ai\u002F) 是一个简化了现代最佳实践下快速、准确神经网络训练过程的库。\n\n\n\n# TensorFlow 开发\n\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_809fa9630256.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## TensorFlow 学习资源\n\n[TensorFlow](https:\u002F\u002Fwww.tensorflow.org) 是一个端到端的开源机器学习平台。它拥有全面且灵活的工具、库和社区资源生态系统，使研究人员能够推动机器学习领域的前沿发展，也让开发者可以轻松构建和部署基于机器学习的应用程序。\n\n[TensorFlow 入门](https:\u002F\u002Fwww.tensorflow.org\u002Flearn)\n\n[TensorFlow 教程](https:\u002F\u002Fwww.tensorflow.org\u002Ftutorials\u002F)\n\n[TensorFlow 开发者认证 | TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002Fcertificate)\n\n[TensorFlow 社区](https:\u002F\u002Fwww.tensorflow.org\u002Fcommunity\u002F)\n\n[TensorFlow 模型与数据集](https:\u002F\u002Fwww.tensorflow.org\u002Fresources\u002Fmodels-datasets)\n\n[TensorFlow Cloud](https:\u002F\u002Fwww.tensorflow.org\u002Fcloud)\n\n[机器学习教育资源 | TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002Fresources\u002Flearn-ml)\n\n[Coursera 上的最佳 TensorFlow 在线课程](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=tensorflow)\n\n[Udemy 上的最佳 TensorFlow 在线课程](https:\u002F\u002Fwww.udemy.com\u002Fcourses\u002Fsearch\u002F?src=ukw&q=tensorflow)\n\n[使用 TensorFlow 进行深度学习 | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fdeep-learning-with-tensorflow-certification-training\u002F)\n\n[使用 TensorFlow 进行深度学习 | edX](https:\u002F\u002Fwww.edx.org\u002Fcourse\u002Fdeep-learning-with-tensorflow)\n\n[深度学习中的 TensorFlow 入门 | Udacity ](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fintro-to-tensorflow-for-deep-learning--ud187)\n\n[TensorFlow 入门：机器学习速成课程 | Google Developers](https:\u002F\u002Fdevelopers.google.com\u002Fmachine-learning\u002Fcrash-course\u002Ffirst-steps-with-tensorflow\u002Ftoolkit)\n\n[在 Azure 机器学习中训练并部署 TensorFlow 模型](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fmachine-learning\u002Fhow-to-train-tensorflow)\n\n[在 Azure Functions 中使用 Python 和 TensorFlow 部署机器学习模型 | Microsoft Azure](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fazure-functions\u002Ffunctions-machine-learning-tensorflow?tabs=bash)\n\n[使用 TensorFlow 进行深度学习 | 亚马逊云服务 (AWS)](https:\u002F\u002Faws.amazon.com\u002Ftensorflow\u002F)\n\n[TensorFlow - Amazon EMR | AWS 文档](https:\u002F\u002Fdocs.aws.amazon.com\u002Femr\u002Flatest\u002FReleaseGuide\u002Femr-tensorflow.html)\n\n[TensorFlow Enterprise | Google Cloud](https:\u002F\u002Fcloud.google.com\u002Ftensorflow-enterprise\u002F)\n\n## TensorFlow 工具、库和框架\n\n[TensorFlow Lite](https:\u002F\u002Fwww.tensorflow.org\u002Flite\u002F) 是一个开源的深度学习框架，用于在移动设备和物联网设备上部署机器学习模型。\n\n[TensorFlow.js](https:\u002F\u002Fwww.tensorflow.org\u002Fjs) 是一个 JavaScript 库，允许你使用 JavaScript 开发或运行机器学习模型，并直接在浏览器客户端、通过 Node.js 在服务器端、通过 React Native 在移动端、通过 Electron 在桌面端，甚至通过 Node.js 在树莓派等物联网设备上使用机器学习。\n\n[Tensorflow_macOS](https:\u002F\u002Fgithub.com\u002Fapple\u002Ftensorflow_macos) 是针对 macOS 11.0 及以上版本优化的 TensorFlow 及其扩展库版本，利用 Apple 的 ML Compute 框架进行加速。\n\n[Google Colaboratory](https:\u002F\u002Fcolab.sandbox.google.com\u002Fnotebooks\u002Fwelcome.ipynb) 是一个免费的 Jupyter 笔记本环境，无需任何配置即可在云端运行，让你只需点击一下就能在浏览器中执行 TensorFlow 代码。\n\n[What-If Tool](https:\u002F\u002Fpair-code.github.io\u002Fwhat-if-tool\u002F) 是一款无需编码即可对机器学习模型进行探查的工具，有助于理解、调试和确保模型公平性。该工具可在 TensorBoard 或 Jupyter\u002FColab 笔记本中使用。\n\n[TensorBoard](https:\u002F\u002Fwww.tensorflow.org\u002Ftensorboard) 是一套可视化工具，用于理解、调试和优化 TensorFlow 程序。\n\n[Keras](https:\u002F\u002Fkeras.io) 是一个高级神经网络 API，用 Python 编写，可在 TensorFlow、CNTK 或 Theano 之上运行。它专注于支持快速实验，也可以在 TensorFlow、Microsoft Cognitive Toolkit、R、Theano 或 PlaidML 等框架之上运行。\n\n[XLA（加速线性代数）](https:\u002F\u002Fwww.tensorflow.org\u002Fxla) 是一个针对线性代数领域的专用编译器，用于优化 TensorFlow 的计算。使用 XLA 后，TensorFlow 在服务器端和移动设备上的速度、内存使用和可移植性都会得到提升。\n\n[ML Perf](https:\u002F\u002Fmlperf.org\u002F) 是一个广泛的机器学习基准测试套件，用于衡量机器学习软件框架、硬件加速器和云平台的性能。\n\n[TensorFlow Playground](https:\u002F\u002Fplayground.tensorflow.org\u002F#activation=tanh&batchSize=10&dataset=circle&regDataset=reg-plane&learningRate=0.03&regularizationRate=0&noise=0&networkShape=4,2&seed=0.04620&showTestData=false&discretize=false&percTrainData=50&x=true&y=true&xTimesY=false&xSquared=false&ySquared=false&cosX=false&sinX=false&cosY=false&sinY=false&collectStats=false&problem=classification&initZero=false&hideText=false) 是一个在浏览器中玩转神经网络的开发环境。\n\n[TPU 研究云 (TRC)](https:\u002F\u002Fsites.research.google\u002Ftrc\u002F) 是一项计划，允许研究人员申请免费使用由 1000 多个 Cloud TPU 组成的集群，以帮助他们加速下一波研究突破。\n\n[MLIR](https:\u002F\u002Fwww.tensorflow.org\u002Fmlir) 是一种新的中间表示和编译器框架。\n\n[Lattice](https:\u002F\u002Fwww.tensorflow.org\u002Flattice) 是一个库，用于构建灵活、可控且可解释的机器学习解决方案，并施加符合常识的形状约束。\n\n[TensorFlow Hub](https:\u002F\u002Fwww.tensorflow.org\u002Fhub) 是一个可重用机器学习模型的库。只需少量代码，即可下载并复用最新的训练好的模型。\n\n[TensorFlow Cloud](https:\u002F\u002Fwww.tensorflow.org\u002Fcloud) 是一个将本地环境连接到 Google Cloud 的库。\n\n[TensorFlow 模型优化工具包](https:\u002F\u002Fwww.tensorflow.org\u002Fmodel_optimization) 是一套用于优化机器学习模型以便部署和执行的工具。\n\n[TensorFlow 推荐系统](https:\u002F\u002Fwww.tensorflow.org\u002Frecommenders) 是一个用于构建推荐系统模型的库。\n\n[TensorFlow Text](https:\u002F\u002Fwww.tensorflow.org\u002Ftext) 是一系列与文本和自然语言处理相关的类和操作，可直接与 TensorFlow 2 配合使用。\n\n[TensorFlow Graphics](https:\u002F\u002Fwww.tensorflow.org\u002Fgraphics) 是一个包含计算机图形功能的库，涵盖相机、灯光、材质以及渲染器等。\n\n[TensorFlow Federated](https:\u002F\u002Fwww.tensorflow.org\u002Ffederated) 是一个开源框架，用于在去中心化数据上进行机器学习和其他计算。\n\n[TensorFlow Probability](https:\u002F\u002Fwww.tensorflow.org\u002Fprobability) 是一个用于概率推理和统计分析的库。\n\n[Tensor2Tensor](https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Ftensor2tensor) 是一个深度学习模型和数据集的库，旨在使深度学习更易用并加速机器学习研究。\n\n[TensorFlow Privacy](https:\u002F\u002Fwww.tensorflow.org\u002Fresponsible_ai\u002Fprivacy) 是一个 Python 库，其中包含了用于在差分隐私保护下训练机器学习模型的 TensorFlow 优化器实现。\n\n[TensorFlow Ranking](https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Franking) 是一个在 TensorFlow 平台上实现排序学习（LTR）技术的库。\n\n[TensorFlow Agents](https:\u002F\u002Fwww.tensorflow.org\u002Fagents) 是一个用于 TensorFlow 中强化学习的库。\n\n[TensorFlow Addons](https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Faddons) 是一个遵循成熟 API 规范但实现了核心 TensorFlow 中未提供的新功能的贡献库，由 [SIG Addons](https:\u002F\u002Fgroups.google.com\u002Fa\u002Ftensorflow.org\u002Fg\u002Faddons) 维护。TensorFlow 本身原生支持大量的运算符、层、指标、损失函数和优化器。\n\n[TensorFlow I\u002FO](https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Fio) 是一个由 SIG IO 维护的数据集、流媒体和文件系统扩展库。\n\n[TensorFlow Quantum](https:\u002F\u002Fwww.tensorflow.org\u002Fquantum) 是一个量子机器学习库，用于快速原型化混合量子-经典机器学习模型。\n\n[Dopamine](https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fdopamine) 是一个用于快速原型化强化学习算法的研究框架。\n\n[TRFL](https:\u002F\u002Fdeepmind.com\u002Fblog\u002Ftrfl\u002F) 是 DeepMind 创建的一个用于强化学习构建模块的库。\n\n[Mesh TensorFlow](https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Fmesh) 是一种用于分布式深度学习的语言，能够指定广泛的分布式张量计算任务。\n\n[RaggedTensors](https:\u002F\u002Fwww.tensorflow.org\u002Fguide\u002Fragged_tensor) 是一个 API，可方便地存储和操作具有非均匀形状的数据，包括文本（单词、句子、字符）以及长度可变的批次。\n\n[Unicode Ops](https:\u002F\u002Fwww.tensorflow.org\u002Ftutorials\u002Fload_data\u002Funicode) 是一个 API，支持在 TensorFlow 中直接处理 Unicode 文本。\n\n[Magenta](https:\u002F\u002Fmagenta.tensorflow.org\u002F) 是一个研究项目，探索机器学习在艺术和音乐创作过程中的作用。\n\n[Nucleus](https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fnucleus) 是一个由 Python 和 C++ 代码组成的库，旨在简化对 SAM 和 VCF 等常见基因组学文件格式的数据读取、写入和分析工作。\n\n[Sonnet](https:\u002F\u002Fgithub.com\u002Fdeepmind\u002Fsonnet) 是 DeepMind 提供的一个用于构建神经网络的库。\n\n[Neural Structured Learning](https:\u002F\u002Fwww.tensorflow.org\u002Fneural_structured_learning) 是一个学习框架，可在利用特征输入的同时，借助结构化信号来训练神经网络。\n\n[Model Remediation](https:\u002F\u002Fwww.tensorflow.org\u002Fresponsible_ai\u002Fmodel_remediation) 是一个库，旨在帮助以减少或消除因潜在性能偏差而导致用户受损的方式创建和训练模型。\n\n[Fairness Indicators](https:\u002F\u002Fwww.tensorflow.org\u002Fresponsible_ai\u002Ffairness_indicators\u002Fguide) 是一个库，可轻松计算二分类和多分类分类器的常用公平性指标。\n\n[Decision Forests](https:\u002F\u002Fwww.tensorflow.org\u002Fdecision_forests) 是用于训练、服务和解释基于决策森林的分类、回归和排序模型的最先进算法。\n\n\n\n# 核心机器学习开发\n\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_602cc9e929a7.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## Core ML 学习资源\n\n[Core ML](https:\u002F\u002Fdeveloper.apple.com\u002Fdocumentation\u002Fcoreml) 是苹果公司推出的一个框架，用于将机器学习模型集成到运行在苹果设备（包括 iOS、watchOS、macOS 和 tvOS）上的应用程序中。Core ML 引入了一种公开的文件格式 (.mlmodel)，支持多种机器学习方法，包括卷积神经网络和循环神经网络、基于提升法的树 ensemble 模型以及广义线性模型。以该格式存储的模型可以通过 Xcode 直接集成到应用中。\n\n[Core ML 简介](https:\u002F\u002Fcoremltools.readme.io\u002Fdocs)\n\n[将 Core ML 模型集成到您的应用中](https:\u002F\u002Fdeveloper.apple.com\u002Fdocumentation\u002Fcoreml\u002Fintegrating_a_core_ml_model_into_your_app)\n\n[Core ML 模型](https:\u002F\u002Fdeveloper.apple.com\u002Fmachine-learning\u002Fmodels\u002F)\n\n[Core ML API 参考文档](https:\u002F\u002Fapple.github.io\u002Fcoremltools\u002Findex.html)\n\n[Core ML 规范](https:\u002F\u002Fapple.github.io\u002Fcoremltools\u002Fmlmodel\u002Findex.html)\n\n[Apple 开发者论坛：Core ML](https:\u002F\u002Fdeveloper.apple.com\u002Fforums\u002Ftags\u002Fcore-ml)\n\n[线上顶级 Core ML 课程 | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002FCore-ML\u002F)\n\n[线上顶级 Core ML 课程 | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=core%20ml)\n\n[IBM Watson 服务与 Core ML | IBM](https:\u002F\u002Fwww.ibm.com\u002Fwatson\u002Fstories\u002Fcoreml)\n\n[使用 IBM Maximo Visual Inspection 生成 Core ML 资产 | IBM](https:\u002F\u002Fdeveloper.ibm.com\u002Ftechnologies\u002Fiot\u002Ftutorials\u002Fibm-maximo-visual-inspection-apple-devices\u002F)\n\n## Core ML 工具、库和框架\n\n[Core ML 工具](https:\u002F\u002Fcoremltools.readme.io\u002F) 是一个包含用于 Core ML 模型转换、编辑和验证的支持工具的项目。\n\n[Create ML](https:\u002F\u002Fdeveloper.apple.com\u002Fmachine-learning\u002Fcreate-ml\u002F) 是一款工具，可在您的 Mac 上提供训练机器学习模型的新方式。它简化了模型训练的复杂性，同时生成功能强大的 Core ML 模型。\n\n[Tensorflow_macOS](https:\u002F\u002Fgithub.com\u002Fapple\u002Ftensorflow_macos) 是针对 macOS 11.0 及更高版本优化的 TensorFlow 和 TensorFlow Addons 版本，并使用 Apple 的 ML Compute 框架进行加速。\n\n[Apple Vision](https:\u002F\u002Fdeveloper.apple.com\u002Fdocumentation\u002Fvision) 是一个框架，可执行人脸和面部特征点检测、文本检测、条形码识别、图像配准以及通用特征跟踪等功能。Vision 还允许使用自定义 Core ML 模型来完成分类或目标检测等任务。\n\n[Keras](https:\u002F\u002Fkeras.io) 是一个用 Python 编写的高级神经网络 API，可在 TensorFlow、CNTK 或 Theano 等后端上运行。它专注于支持快速实验，能够运行在 TensorFlow、Microsoft Cognitive Toolkit、R、Theano 或 PlaidML 等后端之上。\n\n[XGBoost](https:\u002F\u002Fxgboost.readthedocs.io\u002F) 是一个优化的分布式梯度提升库，设计高效、灵活且可移植。它实现了梯度提升框架下的机器学习算法。XGBoost 提供了一种并行树提升（也称为 GBDT、GBM），能够以快速且准确的方式解决许多数据科学问题。它支持在多台机器上进行分布式训练，包括 AWS、GCE、Azure 和 Yarn 集群。此外，它还可以与 Flink、Spark 和其他云数据流系统集成。\n\n[LIBSVM](https:\u002F\u002Fwww.csie.ntu.edu.tw\u002F~cjlin\u002Flibsvm\u002F) 是一套用于支持向量分类（C-SVC、nu-SVC）、回归（epsilon-SVR、nu-SVR）和分布估计（单类 SVM）的集成软件。它支持多类分类。\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) 是一种简单而高效的工具，适用于数据挖掘和数据分析。它基于 NumPy、SciPy 和 Matplotlib 构建。\n\n[Xcode](https:\u002F\u002Fdeveloper.apple.com\u002Fxcode\u002F) 包含开发者为 Mac、iPhone、iPad、Apple TV 和 Apple Watch 创建优秀应用程序所需的一切。Xcode 为开发者提供了一个统一的工作流程，用于用户界面设计、编码、测试和调试。Xcode 是一款通用应用程序，在基于 Intel 的 CPU 和 Apple Silicon 上均能 100% 原生运行。它包含一个统一的 macOS SDK，其中包含了构建可在 Apple Silicon 和 Intel x86_64 CPU 上原生运行的应用程序所需的所有框架、编译器、调试器和其他工具。\n\n[SwiftUI](https:\u002F\u002Fdeveloper.apple.com\u002Fdocumentation\u002Fswiftui) 是一个用户界面工具包，提供视图、控件和布局结构，用于声明应用的用户界面。SwiftUI 框架还提供了事件处理程序，用于将点击、手势和其他类型的输入传递给您的应用程序。\n\n[UIKit](https:\u002F\u002Fdeveloper.apple.com\u002Fdocumentation\u002Fuikit) 是一个框架，为您的 iOS 或 tvOS 应用程序提供必要的基础设施。它提供了窗口和视图架构，用于实现您的界面；事件处理基础设施，用于将多点触控和其他类型的输入传递给您的应用；以及管理用户、系统和您的应用之间交互所需的主运行循环。\n\n[AppKit](https:\u002F\u002Fdeveloper.apple.com\u002Fdocumentation\u002Fappkit) 是一个图形用户界面工具包，包含实现 macOS 应用程序用户界面所需的所有对象，例如窗口、面板、按钮、菜单、滚动条和文本字段。它会为您处理所有细节，例如高效地在屏幕上绘制内容、与硬件设备和屏幕缓冲区通信、在绘制前清除屏幕区域以及裁剪视图等。\n\n[ARKit](https:\u002F\u002Fdeveloper.apple.com\u002Faugmented-reality\u002Farkit\u002F) 是一组由 Apple 开发的软件开发工具，旨在帮助开发者为 iOS 平台构建增强现实应用程序。最新版本 ARKit 3.5 利用 iPad Pro（2020）上的全新 LiDAR 扫描仪和深度感知系统，支持新一代 AR 应用程序，这些应用利用场景几何信息来增强场景理解和物体遮挡效果。\n\n[RealityKit](https:\u002F\u002Fdeveloper.apple.com\u002Fdocumentation\u002Frealitykit) 是一个框架，可利用 ARKit 框架提供的信息实现高性能的 3D 仿真和渲染，从而将虚拟对象无缝融入现实世界。\n\n[SceneKit](https:\u002F\u002Fdeveloper.apple.com\u002Fscenekit\u002F) 是一个高级 3D 图形框架，可帮助您在 iOS 应用程序中创建 3D 动画场景和特效。\n\n[Instruments](https:\u002F\u002Fhelp.apple.com\u002Finstruments\u002Fmac\u002Fcurrent\u002F#\u002Fdev7b09c84f5) 是 Xcode 工具集中的一个强大而灵活的性能分析和测试工具。它旨在帮助您对 iOS、watchOS、tvOS 和 macOS 应用程序、进程和设备进行性能剖析，以便更好地理解并优化其行为和性能。\n\n[Cocoapods](https:\u002F\u002Fcocoapods.org\u002F) 是 Swift 和 Objective-C 的依赖管理器，通过在简单的文本文件中指定项目的依赖关系，即可在 Xcode 项目中使用。CocoaPods 会递归解析各个库之间的依赖关系，获取所有依赖项的源代码，并创建和维护一个 Xcode 工作区来构建您的项目。\n\n[AppCode](https:\u002F\u002Fwww.jetbrains.com\u002Fobjc\u002F) 会持续监控您的代码质量。它会警告您代码中的错误和潜在问题，并自动建议快速修复方案。AppCode 提供大量针对 Objective-C、Swift、C\u002FC++ 的代码检查，以及针对其他受支持语言的多项代码检查。\n\n\n# 深度学习开发\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_d9156e28e925.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## 深度学习学习资源\n\n[深度学习](https:\u002F\u002Fwww.ibm.com\u002Fcloud\u002Flearn\u002Fdeep-learning) 是机器学习的一个子集，本质上是一种具有三个或更多层的神经网络。这些神经网络试图模拟人脑的行为，尽管远未达到人脑的能力水平。这使得神经网络能够从大量数据中“学习”。学习方式可以是[监督学习](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FSupervised_learning)、[半监督学习](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FSemi-supervised_learning)或[无监督学习](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FUnsupervised_learning)。\n\n[NVIDIA深度学习在线课程 | NVIDIA](https:\u002F\u002Fwww.nvidia.com\u002Fen-us\u002Ftraining\u002Fonline\u002F)\n\n[顶级深度学习在线课程 | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=deep%20learning)\n\n[顶级深度学习在线课程 | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fdeep-learning\u002F)\n\n[通过在线课程和教程学习深度学习 | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fdeep-learning)\n\n[Udacity深度学习在线课程纳米学位 | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fdeep-learning-nanodegree--nd101)\n\n[Andrew Ng的机器学习课程 | Coursera](https:\u002F\u002Fwww.coursera.org\u002Flearn\u002Fmachine-learning?)\n\n[Andrew Ng的生产级机器学习工程（MLOps）课程 | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fspecializations\u002Fmachine-learning-engineering-for-production-mlops)\n\n[数据科学：Python中的深度学习与神经网络 | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fdata-science-deep-learning-in-python\u002F)\n\n[使用Python理解机器学习 | Pluralsight](https:\u002F\u002Fwww.pluralsight.com\u002Fcourses\u002Fpython-understanding-machine-learning)\n\n[如何思考机器学习算法 | Pluralsight](https:\u002F\u002Fwww.pluralsight.com\u002Fcourses\u002Fmachine-learning-algorithms)\n\n[深度学习课程 | 斯坦福大学在线](https:\u002F\u002Fonline.stanford.edu\u002Fcourses\u002Fcs230-deep-learning)\n\n[深度学习——华盛顿大学专业与继续教育](https:\u002F\u002Fwww.pce.uw.edu\u002Fcourses\u002Fdeep-learning)\n\n[哈佛大学深度学习在线课程 | 哈佛大学在线学习](https:\u002F\u002Fonline-learning.harvard.edu\u002Fcourse\u002Fdeep-learning-0)\n\n[面向所有人的机器学习课程 | DataCamp](https:\u002F\u002Fwww.datacamp.com\u002Fcourses\u002Fintroduction-to-machine-learning-with-r)\n\n[人工智能专家课程：白金版 | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fartificial-intelligence-exposed-future-10-extreme-edition\u002F)\n\n[顶级人工智能在线课程 | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=artificial%20intelligence)\n\n[通过在线课程和教程学习人工智能 | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fartificial-intelligence)\n\n[人工智能计算机科学专业证书 | edX](https:\u002F\u002Fwww.edx.org\u002Fprofessional-certificate\u002Fharvardx-computer-science-for-artifical-intelligence)\n\n[人工智能纳米学位项目](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fai-artificial-intelligence-nanodegree--nd898)\n\n[Udacity人工智能在线课程 | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fschool-of-ai)\n\n[人工智能入门课程 | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fintro-to-artificial-intelligence--cs271)\n\n[面向物联网开发者的边缘AI课程 | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fintel-edge-ai-for-iot-developers-nanodegree--nd131)\n\n[推理：目标树与基于规则的专家系统 | MIT开放课程资源](https:\u002F\u002Focw.mit.edu\u002Fcourses\u002Felectrical-engineering-and-computer-science\u002F6-034-artificial-intelligence-fall-2010\u002Flecture-videos\u002Flecture-3-reasoning-goal-trees-and-rule-based-expert-systems\u002F)\n\n[专家系统与应用人工智能](https:\u002F\u002Fwww.umsl.edu\u002F~joshik\u002Fmsis480\u002Fchapt11.htm)\n\n[自主系统——微软人工智能](https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002Fai\u002Fautonomous-systems)\n\n[Microsoft Project Bonsai简介](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Flearn\u002Fautonomous-systems\u002Fintro-to-project-bonsai\u002F)\n\n[使用微软自主系统平台进行机器教学](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Farchitecture\u002Fsolution-ideas\u002Farticles\u002Fautonomous-systems)\n\n[自主海事系统培训 | AMC Search](https:\u002F\u002Fwww.amcsearch.com.au\u002Fams-training)\n\n[顶级自动驾驶汽车在线课程 | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fautonomous-cars\u002F)\n\n[应用控制系统1：自动驾驶汽车：数学+PID+MPC | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fapplied-systems-control-for-engineers-modelling-pid-mpc\u002F)\n\n[通过在线课程和教程学习自主机器人技术 | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fautonomous-robotics)\n\n[人工智能纳米学位项目](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fai-artificial-intelligence-nanodegree--nd898)\n\n[Udacity自主系统在线课程及项目 | Udacity自主系统学院](https:\u002F\u002Fwww.udacity.com\u002Fschool-of-autonomous-systems)\n\n[面向物联网开发者的边缘AI课程 | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fintel-edge-ai-for-iot-developers-nanodegree--nd131)\n\n[自主系统MOOC及免费在线课程 | MOOC List](https:\u002F\u002Fwww.mooc-list.com\u002Ftags\u002Fautonomous-systems)\n\n[斯坦福大学在线机器人与自主系统研究生项目](https:\u002F\u002Fonline.stanford.edu\u002Fprograms\u002Frobotics-and-autonomous-systems-graduate-program)\n\n[移动自主系统实验室 | MIT开放课程资源](https:\u002F\u002Focw.mit.edu\u002Fcourses\u002Felectrical-engineering-and-computer-science\u002F6-186-mobile-autonomous-systems-laboratory-january-iap-2005\u002Flecture-notes\u002F)\n\n## 深度学习工具、库和框架\n\n[NVIDIA cuDNN](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcudnn) 是一个用于[深度神经网络](https:\u002F\u002Fdeveloper.nvidia.com\u002Fdeep-learning)的GPU加速原语库。cuDNN为前向和反向卷积、池化、归一化以及激活层等标准操作提供了高度优化的实现。cuDNN可以加速广泛使用的深度学习框架，包括[Caffe2](https:\u002F\u002Fcaffe2.ai\u002F)、[Chainer](https:\u002F\u002Fchainer.org\u002F)、[Keras](https:\u002F\u002Fkeras.io\u002F)、[MATLAB](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fdeep-learning.html)、[MxNet](https:\u002F\u002Fmxnet.incubator.apache.org\u002F)、[PyTorch](https:\u002F\u002Fpytorch.org\u002F)和[TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002F)。\n\n[NVIDIA DLSS（深度学习超级采样）](https:\u002F\u002Fdeveloper.nvidia.com\u002Fdlss)是一种基于时间的图像超分辨率AI渲染技术，它利用GeForce RTX™ GPU上的专用Tensor Core AI处理器来提升图形性能。DLSS借助深度学习神经网络的强大功能，可提高帧率并为您的游戏生成清晰锐利的画面。\n\n[AMD FidelityFX 超分辨率（FSR）](https:\u002F\u002Fwww.amd.com\u002Fen\u002Ftechnologies\u002Fradeon-software-fidelityfx) 是一种开源、高质量的解决方案，能够从低分辨率输入生成高分辨率帧。它采用了一系列前沿的深度学习算法，尤其注重边缘质量的提升，与直接以原生分辨率渲染相比，可带来显著的性能提升。FSR 为一些计算成本较高的渲染任务（如 AMD RDNA™ 和 AMD RDNA™ 2 架构中的硬件光线追踪）提供了“实用性能”。\n\n[Intel Xe 超采样（XeSS）](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=Y9hfpf-SqEg) 是一种基于时间的图像超分辨率 AI 渲染技术，能够提升图形性能，类似于 NVIDIA 的 [DLSS（深度学习超级采样）](https:\u002F\u002Fdeveloper.nvidia.com\u002Fdlss)。Intel 的 Arc GPU 架构（2022 年初）将配备专门用于运行 XeSS 的 Xe 核心。这些 GPU 还拥有 Xe Matrix eXtenstions 矩阵单元（XMX），用于硬件加速的 AI 处理。尽管如此，XeSS 也可以在没有 XMX 的设备上运行，包括集成显卡，但其在非 Intel 显卡上的性能会较低，因为届时将依赖 [DP4a 指令](https:\u002F\u002Fwww.intel.com\u002Fcontent\u002Fdam\u002Fwww\u002Fpublic\u002Fus\u002Fen\u002Fdocuments\u002Freference-guides\u002F11th-gen-quick-reference-guide.pdf)。\n\n[Jupyter Notebook](https:\u002F\u002Fjupyter.org\u002F) 是一款开源 Web 应用程序，允许用户创建和共享包含实时代码、公式、可视化内容及叙述性文本的文档。Jupyter 广泛应用于数据清洗与转换、数值模拟、统计建模、数据可视化、数据科学和机器学习等领域。\n\n[Apache Spark](https:\u002F\u002Fspark.apache.org\u002F) 是一个用于大规模数据处理的统一分析引擎。它提供 Scala、Java、Python 和 R 等高级 API，并配备优化的执行引擎，支持通用计算图进行数据分析。此外，Spark 还提供丰富的高级工具集，包括用于 SQL 和 DataFrame 的 Spark SQL、用于机器学习的 MLlib、用于图计算的 GraphX 以及用于流式处理的 Structured Streaming。\n\n[Apache Spark Connector for SQL Server and Azure SQL](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fsql-spark-connector) 是一款高性能连接器，使您能够在大数据分析中使用事务型数据，并将结果持久化以供即席查询或报告使用。该连接器允许您将任何 SQL 数据库——无论是本地部署还是云端——用作 Spark 作业的输入数据源或输出数据接收端。\n\n[Apache PredictionIO](https:\u002F\u002Fpredictionio.apache.org\u002F) 是一个面向开发者、数据科学家和最终用户的开源机器学习框架。它支持事件收集、算法部署、模型评估以及通过 REST API 查询预测结果。PredictionIO 基于 Hadoop、HBase（以及其他数据库）、Elasticsearch 和 Spark 等可扩展的开源服务，并实现了所谓的 Lambda 架构。\n\n[Cluster Manager for Apache Kafka (CMAK)](https:\u002F\u002Fgithub.com\u002Fyahoo\u002FCMAK) 是一款用于管理 [Apache Kafka](https:\u002F\u002Fkafka.apache.org\u002F) 集群的工具。\n\n[BigDL](https:\u002F\u002Fbigdl-project.github.io\u002F) 是一个专为 Apache Spark 设计的分布式深度学习库。借助 BigDL，用户可以将深度学习应用程序编写为标准的 Spark 程序，这些程序可以直接在现有的 Spark 或 Hadoop 集群上运行。\n\n[Eclipse Deeplearning4J (DL4J)](https:\u002F\u002Fdeeplearning4j.konduit.ai\u002F) 是一组旨在满足基于 JVM（Scala、Kotlin、Clojure 和 Groovy）的深度学习应用所有需求的项目。这意味着从原始数据开始，无论其来源和格式如何，都可以对其进行加载和预处理，进而构建和调优各种简单或复杂的深度学习网络。\n\n[Deep Learning Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdeep-learning.html) 是一款提供框架的工具，可用于设计和实现深度神经网络，内置算法、预训练模型和应用程序。您可以使用卷积神经网络（ConvNets、CNNs）和长短期记忆网络（LSTM）对图像、时间序列和文本数据进行分类和回归。借助自动微分、自定义训练循环和权重共享，还可以构建生成对抗网络（GANs）和暹罗网络等架构。通过 Deep Network Designer 应用程序，您可以以图形化方式设计、分析和训练网络。该工具可通过 ONNX 格式与 TensorFlow™ 和 PyTorch 交换模型，并可导入来自 TensorFlow-Keras 和 Caffe 的模型。该工具箱还支持迁移学习，兼容 DarkNet-53、ResNet-50、NASNet、SqueezeNet 等多种预训练模型。\n\n[Reinforcement Learning Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Freinforcement-learning.html) 是一款提供应用程序、函数和 Simulink® 块的工具，用于利用强化学习算法（包括 DQN、PPO、SAC 和 DDPG）训练策略。您可以使用这些策略来实现控制器和决策算法，适用于资源分配、机器人技术和自主系统等复杂应用场景。\n\n[Deep Learning HDL Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdeep-learning-hdl.html) 是一款提供函数和工具的工具，用于在 FPGA 和 SoC 上原型化并实现深度学习网络。它提供预构建的比特流，可在受支持的 Xilinx® 和 Intel® FPGA 以及 SoC 设备上运行各种深度学习网络。同时，剖析和估算工具可以帮助您通过权衡设计、性能和资源利用率之间的关系，定制深度学习网络。\n\n[Parallel Computing Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmatlab-parallel-server.html) 是一款允许您使用多核处理器、GPU 和计算机集群解决计算密集型及数据密集型问题的工具。高级构造块，如并行 for 循环、特殊数组类型和并行化数值算法，使您无需 CUDA 或 MPI 编程即可实现 MATLAB® 应用程序的并行化。该工具箱允许您在 MATLAB 及其他工具箱中使用支持并行计算的函数。结合 Simulink® 使用时，您可以在同一时间并行运行多个模型仿真。程序和模型既可以在交互模式下运行，也可以以批处理模式运行。\n\n[XGBoost](https:\u002F\u002Fxgboost.readthedocs.io\u002F) 是一个经过优化的分布式梯度提升库，具有高效、灵活和可移植的特点。它在梯度提升框架下实现了机器学习算法。XGBoost 提供一种并行树提升方法（也称为 GBDT 或 GBM），能够快速且准确地解决许多数据科学问题。它支持在多台机器上进行分布式训练，包括 AWS、GCE、Azure 和 Yarn 集群。此外，它还可以与 Flink、Spark 和其他云数据流系统集成。\n\n[LIBSVM](https:\u002F\u002Fwww.csie.ntu.edu.tw\u002F~cjlin\u002Flibsvm\u002F) 是一个用于支持向量分类（C-SVC、nu-SVC）、回归（epsilon-SVR、nu-SVR）以及分布估计（一类 SVM）的集成软件。它支持多类分类。\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) 是一款简单高效的用于数据挖掘和数据分析的工具。它基于 NumPy、SciPy 和 Matplotlib 构建。\n\n[TensorFlow](https:\u002F\u002Fwww.tensorflow.org) 是一个端到端的开源机器学习平台。它拥有全面且灵活的工具、库和社区资源生态系统，使研究人员能够推动机器学习领域的前沿发展，同时也让开发者能够轻松构建和部署基于机器学习的应用程序。\n\n[Keras](https:\u002F\u002Fkeras.io) 是一个用 Python 编写的高级神经网络 API，可运行在 TensorFlow、CNTK 或 Theano 之上。它专注于加速实验，支持在 TensorFlow、Microsoft Cognitive Toolkit、R、Theano 或 PlaidML 等框架之上运行。\n\n[PyTorch](https:\u002F\u002Fpytorch.org) 是一个针对不规则输入数据（如图、点云和流形）的深度学习库，主要由 Facebook 的 AI 研究实验室开发。\n\n[Azure Databricks](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fdatabricks\u002F) 是一种快速且协作式的基于 Apache Spark 的大数据分析服务，专为数据科学和数据工程设计。Azure Databricks 可在几分钟内搭建 Apache Spark 环境，支持自动扩展，并允许用户在交互式工作区中协作处理共享项目。Azure Databricks 支持 Python、Scala、R、Java 和 SQL，以及 TensorFlow、PyTorch 和 scikit-learn 等数据科学框架和库。\n\n[Microsoft Cognitive Toolkit (CNTK)](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fcognitive-toolkit\u002F) 是一个面向商业级分布式深度学习的开源工具包。它通过有向图将神经网络描述为一系列计算步骤。CNTK 允许用户轻松实现并组合常用的模型类型，如前馈 DNN、卷积神经网络（CNN）和循环神经网络（RNN\u002FLSTM）。CNTK 实现了随机梯度下降（SGD，误差反向传播）学习，并支持自动微分及跨多个 GPU 和服务器的并行化。\n\n[Tensorflow_macOS](https:\u002F\u002Fgithub.com\u002Fapple\u002Ftensorflow_macos) 是针对 macOS 11.0 及以上版本优化的 TensorFlow 及 TensorFlow Addons 版本，利用 Apple 的 ML Compute 框架进行加速。\n\n[Apache Airflow](https:\u002F\u002Fairflow.apache.org) 是一个由社区创建的开源工作流管理平台，用于以编程方式编写、调度和监控工作流。安装简便、原则清晰、可扩展性强。Airflow 具有模块化架构，使用消息队列来编排任意数量的工作节点，能够无限扩展。\n\n[Open Neural Network Exchange(ONNX)](https:\u002F\u002Fgithub.com\u002Fonnx) 是一个开放的生态系统，使 AI 开发者能够根据项目的发展选择合适的工具。ONNX 提供了一种开源的 AI 模型格式，适用于深度学习和传统机器学习。它定义了一个可扩展的计算图模型，以及内置算子和标准数据类型的规范。\n\n[Apache MXNet](https:\u002F\u002Fmxnet.apache.org\u002F) 是一个兼顾效率与灵活性的深度学习框架。它允许混合符号式和命令式编程，以最大化效率和生产力。MXNet 的核心是一个动态依赖调度器，可实时自动并行化符号式和命令式操作。在其之上还有一层图优化层，使符号式执行既快速又节省内存。MXNet 具有良好的移植性和轻量化特性，能够有效扩展到多 GPU 和多台机器上。支持 Python、R、Julia、Scala、Go、JavaScript 等多种语言。\n\n[AutoGluon](https:\u002F\u002Fautogluon.mxnet.io\u002Findex.html) 是一个自动化机器学习工具包，能够简化深度学习任务，帮助用户在应用中轻松获得强大的预测性能。只需几行代码，即可在表格数据、图像和文本数据上训练并部署高精度的深度学习模型。\n\n[Anaconda](https:\u002F\u002Fwww.anaconda.com\u002F) 是一个非常流行的数据科学平台，适用于机器学习和深度学习，帮助用户开发、训练和部署模型。\n\n[PlaidML](https:\u002F\u002Fgithub.com\u002Fplaidml\u002Fplaidml) 是一个先进且可移植的张量编译器，旨在使深度学习能够在笔记本电脑、嵌入式设备或其他计算硬件支持不足或软件许可限制较多的设备上运行。\n\n[OpenCV](https:\u002F\u002Fopencv.org) 是一个高度优化的库，专注于实时计算机视觉应用。其 C++、Python 和 Java 接口支持 Linux、MacOS、Windows、iOS 和 Android 系统。\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) 是一个基于 SciPy、NumPy 和 Matplotlib 构建的 Python 机器学习模块，使得许多流行的机器学习算法能够更便捷地被应用和实现。\n\n[Weka](https:\u002F\u002Fwww.cs.waikato.ac.nz\u002Fml\u002Fweka\u002F) 是一款开源的机器学习软件，可通过图形用户界面、标准终端应用程序或 Java API 访问。它广泛应用于教学、科研和工业领域，内置了丰富的工具用于标准机器学习任务，并且还能透明地访问 scikit-learn、R 和 Deeplearning4j 等知名工具箱。\n\n[Caffe](https:\u002F\u002Fgithub.com\u002FBVLC\u002Fcaffe) 是一个以表达性、速度和模块化为核心设计的深度学习框架。它由伯克利人工智能研究实验室（BAIR）、伯克利视觉与学习中心（BVLC）以及社区贡献者共同开发。\n\n[Theano](https:\u002F\u002Fgithub.com\u002FTheano\u002FTheano) 是一个 Python 库，允许高效地定义、优化和评估涉及多维数组的数学表达式，并与 NumPy 紧密集成。\n\n[Microsoft Project Bonsai](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fproject-bonsai\u002F) 是一个低代码 AI 平台，可加速基于 AI 的自动化开发，属于 Microsoft 自动化系统套件的一部分。Bonsai 用于构建能够提供操作指导或自主决策的 AI 组件，以优化工艺参数、提高生产效率并减少停机时间。\n\n[Microsoft AirSim](https:\u002F\u002Fmicrosoft.github.io\u002FAirSim\u002Flidar.html) 是一款基于虚幻引擎（并提供实验性 Unity 版本）的无人机、汽车等模拟器。AirSim 是开源且跨平台的，支持与 PX4 和 ArduPilot 等主流飞行控制器进行【软件在环仿真】（SIL），以及与 PX4 进行【硬件在环】（HIL），从而实现物理和视觉上高度逼真的仿真效果。它以虚幻引擎插件的形式开发，可轻松集成到任何虚幻引擎场景中。AirSim 正在被打造为一个用于人工智能研究的平台，旨在试验深度学习、计算机视觉和强化学习算法在自动驾驶车辆中的应用。\n\n[CARLA](https:\u002F\u002Fgithub.com\u002Fcarla-simulator\u002Fcarla) 是一款面向自动驾驶研究的开源模拟器。CARLA 从零开始设计，专门用于支持自动驾驶系统的开发、训练和验证。除了开放源代码和通信协议外，CARLA 还提供了专为此目的创建的开放数字资产（如城市布局、建筑和车辆），用户可以自由使用这些资源。\n\n[CARLA 的 ROS\u002FROS2 桥接包](https:\u002F\u002Fgithub.com\u002Fcarla-simulator\u002Fros-bridge) 是一个实现 ROS 与 CARLA 之间双向通信的桥梁。CARLA 服务器发送的信息会被转换为 ROS 主题；同样地，ROS 节点之间传递的消息也会被翻译成可在 CARLA 中执行的命令。\n\n[ROS 工具箱](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fros.html) 是一种将 MATLAB® 和 Simulink® 与机器人操作系统（ROS 和 ROS 2）连接起来的工具，允许用户构建 ROS 节点网络。该工具箱包含 MATLAB 函数和 Simulink 模块，用于导入、分析和回放以 rosbag 文件格式记录的 ROS 数据。此外，还可以连接到实时 ROS 网络以访问 ROS 消息。\n\n[机器人技术工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Frobotics.html) 提供了一系列专为机器人技术设计的功能（包括机械臂、移动机器人和人形机器人的设计、仿真和测试），充分利用了 MATLAB 的原生能力（线性代数、跨平台性和图形化功能）。该工具箱还特别支持移动机器人，提供了自行车模型等运动学模型、路径规划算法（如虫子算法、距离变换、D*、PRM）、动力学规划（如格网法、RRT）、定位（EKF、粒子滤波器）、地图构建（EKF）以及同时定位与建图（EKF）等功能，并包含一个非完整约束车辆的 Simulink 模型。此外，该工具箱还提供了一个详细的四旋翼飞行机器人 Simulink 模型。\n\n[图像处理工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fimage.html) 提供了一套全面的标准算法和工作流应用程序，用于图像处理、分析、可视化及算法开发。用户可以进行图像分割、图像增强、去噪、几何变换、图像配准以及三维图像处理等操作。\n\n[计算机视觉工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fcomputer-vision.html) 提供用于设计和测试计算机视觉、三维视觉和视频处理系统的算法、函数和应用程序。用户可以执行目标检测与跟踪、特征检测、提取及匹配等任务。此外，还可以自动化单目、双目和鱼眼相机的标定流程。对于三维视觉，该工具箱支持视觉 SLAM 和点云 SLAM、立体视觉、运动恢复结构以及点云处理等功能。\n\n[机器人技术工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Frobotics.html) 是一种将机器人技术相关功能（如机械臂、移动机器人和人形机器人的设计、仿真和测试）引入 MATLAB 的工具箱，充分利用了 MATLAB 的原生优势（线性代数、跨平台性和图形化功能）。该工具箱还特别支持移动机器人，提供了自行车模型等运动学模型、路径规划算法（如虫子算法、距离变换、D*、PRM）、动力学规划（如格网法、RRT）、定位（EKF、粒子滤波器）、地图构建（EKF）以及同时定位与建图（EKF）等功能，并包含一个非完整约束车辆的 Simulink 模型。此外，该工具箱还提供了一个详细的四旋翼飞行机器人 Simulink 模型。\n\n[模型预测控制工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmodel-predictive-control.html) 提供用于设计和仿真线性和非线性模型预测控制器（MPC）的函数、应用程序和 Simulink® 模块。该工具箱允许用户指定被控对象和干扰模型、预测时域、约束条件以及权重。通过运行闭环仿真，用户可以评估控制器的性能。\n\n[预测性维护工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fpredictive-maintenance.html) 允许用户管理传感器数据、设计状态指标，并估算设备的剩余使用寿命（RUL）。该工具箱提供函数和交互式应用程序，用于基于数据和基于模型的技术（包括统计分析、频谱分析和时间序列分析）来探索、提取和排序特征。\n\n[Vision HDL 工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fvision-hdl.html) 提供用于在 FPGA 和 ASIC 上设计与实现视觉系统的像素流算法。它配备了一个设计框架，支持多种接口类型、帧尺寸和帧率。该工具箱中的图像处理、视频和计算机视觉算法采用了适合 HDL 实现的架构。\n\n[自动驾驶工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fautomated-driving.html) 是一种 MATLAB 工具箱，提供用于设计、仿真和测试 ADAS 及自动驾驶系统的算法和工具。用户可以设计和测试视觉与激光雷达感知系统，以及传感器融合、路径规划和车辆控制器等功能。可视化工具包括鸟瞰图、传感器覆盖范围、检测结果和轨迹的示意图，以及用于显示视频、激光雷达数据和地图的界面。该工具箱支持导入并使用 HERE HD Live Map 数据和 OpenDRIVE® 道路网络。此外，它还提供了常见 ADAS 和自动驾驶功能的参考应用示例，包括前向碰撞预警（FCW）、自动紧急制动（AEB）、自适应巡航控制（ACC）、车道保持辅助（LKA）以及代客泊车等功能。该工具箱支持 C\u002FC++ 代码生成，以实现快速原型开发和 HIL 测试，并对传感器融合、目标跟踪、路径规划和车辆控制器算法提供支持。\n\n[UAV工具箱](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fuav.html) 是一款应用程序，提供用于设计、仿真、测试和部署无人机（UAV）及无人飞行器应用的工具和参考应用。您可以设计自主飞行算法、无人机任务以及飞行控制器。飞行日志分析器应用程序允许您以交互方式分析常见飞行日志格式中的三维飞行轨迹、遥测信息和传感器读数。\n\n[导航工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fnavigation.html) 是一种工具，提供用于运动规划、同时定位与地图构建（SLAM）以及惯性导航的算法和分析工具。该工具箱包含可自定义的基于搜索和采样的路径规划器，以及用于验证和比较路径的度量标准。您可以创建二维和三维地图表示，使用SLAM算法生成地图，并通过SLAM地图构建器应用程序以交互方式可视化和调试地图生成过程。\n\n[Lidar工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Flidar.html) 是一种工具，提供用于设计、分析和测试激光雷达处理系统的算法、函数和应用程序。您可以执行目标检测与跟踪、语义分割、形状拟合、激光雷达配准以及障碍物检测等操作。Lidar工具箱支持激光雷达与相机的交叉标定，适用于结合计算机视觉和激光雷达处理的工作流程。\n\n[地图制作工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmapping.html) 是一种工具，提供用于转换地理数据和创建地图显示的算法和函数。您可以在地理背景下可视化数据，利用超过60种地图投影构建地图显示，并将来自各种来源的数据转换为一致的地理坐标系统。\n\n\n\n\n# 强化学习开发\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_d7c013b9339b.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## 强化学习学习资源\n\n[强化学习](https:\u002F\u002Fwww.ibm.com\u002Fcloud\u002Flearn\u002Fdeep-learning#toc-deep-learn-md_Q_Of3) 是机器学习的一个子集，属于具有三层或更多层的神经网络。这些神经网络试图模拟人脑的行为，尽管距离真正媲美人脑的能力还很遥远。通过这一过程，神经网络能够“学习”：模型根据反馈在环境中执行动作，并不断优化以最大化奖励。学习方式可以是[监督学习](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FSupervised_learning)、[半监督学习](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FSemi-supervised_learning)或[无监督学习](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FUnsupervised_learning)。\n\n[顶级强化学习课程 | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=reinforcement%20learning)\n\n[顶级强化学习课程 | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Freinforcement-learning\u002F)\n\n[顶级强化学习课程 | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Freinforcement-learning--ud600)\n\n[强化学习课程 | 斯坦福在线](https:\u002F\u002Fonline.stanford.edu\u002Fcourses\u002Fxcs234-reinforcement-learning)\n\n[深度学习在线课程 | NVIDIA](https:\u002F\u002Fwww.nvidia.com\u002Fen-us\u002Ftraining\u002Fonline\u002F)\n\n[顶级深度学习在线课程 | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=deep%20learning)\n\n[顶级深度学习在线课程 | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fdeep-learning\u002F)\n\n[通过在线课程和教程学习深度学习 | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fdeep-learning)\n\n[深度学习在线课程纳米学位 | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fdeep-learning-nanodegree--nd101)\n\n[吴恩达的机器学习课程 | Coursera](https:\u002F\u002Fwww.coursera.org\u002Flearn\u002Fmachine-learning?)\n\n[吴恩达的生产级机器学习工程（MLOps）课程 | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fspecializations\u002Fmachine-learning-engineering-for-production-mlops)\n\n[数据科学：Python中的深度学习与神经网络 | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fdata-science-deep-learning-in-python\u002F)\n\n[用Python理解机器学习 | Pluralsight](https:\u002F\u002Fwww.pluralsight.com\u002Fcourses\u002Fpython-understanding-machine-learning)\n\n[如何思考机器学习算法 | Pluralsight](https:\u002F\u002Fwww.pluralsight.com\u002Fcourses\u002Fmachine-learning-algorithms)\n\n[深度学习课程 | 斯坦福在线](https:\u002F\u002Fonline.stanford.edu\u002Fcourses\u002Fcs230-deep-learning)\n\n[深度学习——华盛顿大学继续教育项目](https:\u002F\u002Fwww.pce.uw.edu\u002Fcourses\u002Fdeep-learning)\n\n[哈佛大学深度学习在线课程 | 哈佛在线学习](https:\u002F\u002Fonline-learning.harvard.edu\u002Fcourse\u002Fdeep-learning-0)\n\n[面向所有人的机器学习课程 | DataCamp](https:\u002F\u002Fwww.datacamp.com\u002Fcourses\u002Fintroduction-to-machine-learning-with-r)\n\n[人工智能专家课程：白金版 | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fartificial-intelligence-exposed-future-10-extreme-edition\u002F)\n\n[顶级人工智能在线课程 | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=artificial%20intelligence)\n\n[通过在线课程和教程学习人工智能 | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fartificial-intelligence)\n\n[人工智能计算机科学专业证书 | edX](https:\u002F\u002Fwww.edx.org\u002Fprofessional-certificate\u002Fharvardx-computer-science-for-artifical-intelligence)\n\n[人工智能纳米学位项目](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fai-artificial-intelligence-nanodegree--nd898)\n\n[人工智能（AI）在线课程 | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fschool-of-ai)\n\n[Intro to Artificial Intelligence Course | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fintro-to-artificial-intelligence--cs271)\n\n[面向物联网开发者的边缘AI课程 | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fintel-edge-ai-for-iot-developers-nanodegree--nd131)\n\n[推理：目标树与基于规则的专家系统 | MIT OpenCourseWare](https:\u002F\u002Focw.mit.edu\u002Fcourses\u002Felectrical-engineering-and-computer-science\u002F6-034-artificial-intelligence-fall-2010\u002Flecture-videos\u002Flecture-3-reasoning-goal-trees-and-rule-based-expert-systems\u002F)\n\n[专家系统与应用人工智能](https:\u002F\u002Fwww.umsl.edu\u002F~joshik\u002Fmsis480\u002Fchapt11.htm)\n\n[自主系统——微软人工智能](https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002Fai\u002Fautonomous-systems)\n\n[Microsoft Project Bonsai简介](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Flearn\u002Fautonomous-systems\u002Fintro-to-project-bonsai\u002F)\n\n[使用微软自主系统平台进行机器教学](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Farchitecture\u002Fsolution-ideas\u002Farticles\u002Fautonomous-systems)\n\n[自主海事系统培训 | AMC Search](https:\u002F\u002Fwww.amcsearch.com.au\u002Fams-training)\n\n[顶级自动驾驶汽车在线课程 | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fautonomous-cars\u002F)\n\n[应用控制系统1：自动驾驶汽车——数学+PID+MPC | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fapplied-systems-control-for-engineers-modelling-pid-mpc\u002F)\n\n[通过在线课程和教程学习自主机器人技术 | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fautonomous-robotics)\n\n[人工智能纳米学位项目](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fai-artificial-intelligence-nanodegree--nd898)\n\n[自主系统在线课程与项目 | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fschool-of-autonomous-systems)\n\n[面向物联网开发者的边缘AI课程 | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fintel-edge-ai-for-iot-developers-nanodegree--nd131)\n\n[自主系统MOOC及免费在线课程 | MOOC List](https:\u002F\u002Fwww.mooc-list.com\u002Ftags\u002Fautonomous-systems)\n\n[斯坦福在线机器人与自主系统研究生项目](https:\u002F\u002Fonline.stanford.edu\u002Fprograms\u002Frobotics-and-autonomous-systems-graduate-program)\n\n[移动自主系统实验室 | MIT OpenCourseWare](https:\u002F\u002Focw.mit.edu\u002Fcourses\u002Felectrical-engineering-and-computer-science\u002F6-186-mobile-autonomous-systems-laboratory-january-iap-2005\u002Flecture-notes\u002F)\n\n## 强化学习工具、库和框架\n\n[OpenAI Gym](https:\u002F\u002Fgym.openai.com\u002F) 是一个开源的 Python 库，用于开发和比较强化学习算法。它提供了一个标准 API，用于在学习算法和环境之间进行通信，并包含一组符合该 API 标准的环境。\n\n[ReinforcementLearning.jl](https:\u002F\u002Fjuliareinforcementlearning.org\u002F) 是一套用于在 Julia 中进行强化学习研究的工具。\n\n[Reinforcement Learning Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Freinforcement-learning.html) 是一种工具，提供应用程序、函数以及 Simulink® 模块，用于利用强化学习算法（包括 DQN、PPO、SAC 和 DDPG）训练策略。您可以使用这些策略来实现控制器和决策算法，应用于资源分配、机器人技术和自主系统等复杂场景。\n\n[Amazon SageMaker](https:\u002F\u002Faws.amazon.com\u002Frobomaker\u002F) 是一项完全托管的服务，使每位开发者和数据科学家都能快速构建、训练和部署机器学习（ML）模型。\n\n[AWS RoboMaker](https:\u002F\u002Faws.amazon.com\u002Frobomaker\u002F) 是一项服务，为客户提供完全托管且可扩展的仿真基础设施，用于多机器人仿真以及与仿真中的回归测试相结合的 CI\u002FCD 集成。\n\n[TensorFlow](https:\u002F\u002Fwww.tensorflow.org) 是一个端到端的开源机器学习平台。它拥有全面而灵活的工具、库和社区资源生态系统，使研究人员能够推动机器学习领域的前沿发展，同时也让开发者可以轻松构建和部署基于机器学习的应用程序。\n\n[Keras](https:\u002F\u002Fkeras.io) 是一个用 Python 编写的高级神经网络 API，可在 TensorFlow、CNTK 或 Theano 之上运行。它专注于支持快速实验，目前也支持在 TensorFlow、Microsoft Cognitive Toolkit、R、Theano 或 PlaidML 等框架之上运行。\n\n[PyTorch](https:\u002F\u002Fpytorch.org) 是一个用于处理不规则输入数据（如图、点云和流形）的深度学习库，主要由 Facebook 的 AI 研究实验室开发。\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) 是一款简单高效的用于数据挖掘和数据分析的工具。它基于 NumPy、SciPy 和 Matplotlib 构建。\n\n[NVIDIA cuDNN](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcudnn) 是一个针对 [深度神经网络](https:\u002F\u002Fdeveloper.nvidia.com\u002Fdeep-learning) 的 GPU 加速原语库。cuDNN 提供了高度优化的标准实现，例如前向和反向卷积、池化、归一化以及激活层等操作。cuDNN 可以加速广泛使用的深度学习框架，包括 [Caffe2](https:\u002F\u002Fcaffe2.ai\u002F)、[Chainer](https:\u002F\u002Fchainer.org\u002F)、[Keras](https:\u002F\u002Fkeras.io\u002F)、[MATLAB](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fdeep-learning.html)、[MxNet](https:\u002F\u002Fmxnet.incubator.apache.org\u002F)、[PyTorch](https:\u002F\u002Fpytorch.org\u002F) 和 [TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002F) 等。\n\n[Jupyter Notebook](https:\u002F\u002Fjupyter.org\u002F) 是一个开源 Web 应用程序，允许用户创建和共享包含实时代码、公式、可视化内容及叙述性文本的文档。Jupyter 广泛应用于数据清洗与转换、数值模拟、统计建模、数据可视化、数据科学和机器学习等领域。\n\n[Apache Spark](https:\u002F\u002Fspark.apache.org\u002F) 是一个用于大规模数据处理的统一分析引擎。它提供 Scala、Java、Python 和 R 等高级 API，并配备优化的执行引擎，支持用于数据分析的一般计算图。此外，Spark 还提供丰富的高级工具集，包括用于 SQL 和 DataFrame 的 Spark SQL、用于机器学习的 MLlib、用于图处理的 GraphX 以及用于流式处理的 Structured Streaming。\n\n[Apache Spark Connector for SQL Server and Azure SQL](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fsql-spark-connector) 是一种高性能连接器，使您能够在大数据分析中使用事务性数据，并将结果持久化以供即席查询或报告使用。该连接器允许您将任何 SQL 数据库——无论是在本地还是云端——用作 Spark 作业的输入数据源或输出数据目标。\n\n[Apache PredictionIO](https:\u002F\u002Fpredictionio.apache.org\u002F) 是一个面向开发者、数据科学家和最终用户的开源机器学习框架。它支持事件收集、算法部署、评估以及通过 REST API 查询预测结果。该框架基于 Hadoop、HBase（以及其他数据库）、Elasticsearch 和 Spark 等可扩展的开源服务，并实现了所谓的 Lambda 架构。\n\n[Cluster Manager for Apache Kafka (CMAK)](https:\u002F\u002Fgithub.com\u002Fyahoo\u002FCMAK) 是一个用于管理 [Apache Kafka](https:\u002F\u002Fkafka.apache.org\u002F) 集群的工具。\n\n[BigDL](https:\u002F\u002Fbigdl-project.github.io\u002F) 是一个适用于 Apache Spark 的分布式深度学习库。借助 BigDL，用户可以将深度学习应用程序编写为标准的 Spark 程序，这些程序可以直接在现有的 Spark 或 Hadoop 集群上运行。\n\n[Eclipse Deeplearning4J (DL4J)](https:\u002F\u002Fdeeplearning4j.konduit.ai\u002F) 是一组旨在满足基于 JVM（Scala、Kotlin、Clojure 和 Groovy）的深度学习应用所有需求的项目。这意味着从原始数据开始，无论其来源和格式如何，都可以对其进行加载和预处理，进而构建并调优各种简单和复杂的深度学习网络。\n\n[Deep Learning Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdeep-learning.html) 是一个提供框架的工具，可用于设计和实现深度神经网络，包含算法、预训练模型和应用程序。您可以使用卷积神经网络（ConvNets、CNNs）和长短期记忆网络（LSTM）对图像、时间序列和文本数据进行分类和回归分析。还可以利用自动微分、自定义训练循环和共享权重，构建生成对抗网络（GANs）和暹罗网络等网络架构。借助 Deep Network Designer 应用程序，您可以以图形化方式设计、分析和训练网络。该工具可通过 ONNX 格式与 TensorFlow™ 和 PyTorch 交换模型，并可导入来自 TensorFlow-Keras 和 Caffe 的模型。该工具箱支持迁移学习，可与 DarkNet-53、ResNet-50、NASNet、SqueezeNet 等众多预训练模型配合使用。\n\n[Deep Learning HDL Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdeep-learning-hdl.html) 是一个提供函数和工具的工具，用于在 FPGA 和 SoC 上原型化和实现深度学习网络。它提供了预构建的比特流，可在受支持的 Xilinx® 和 Intel® FPGA 以及 SoC 设备上运行各种深度学习网络。同时，性能分析和估算工具可以帮助您通过权衡设计、性能和资源利用率之间的关系来定制深度学习网络。\n\n[Parallel Computing Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmatlab-parallel-server.html) 是一个工具，可让您利用多核处理器、GPU 和计算机集群来解决计算密集型和数据密集型问题。高级构造，如并行 for 循环、特殊数组类型和并行化的数值算法，使您无需 CUDA 或 MPI 编程即可并行化 MATLAB® 应用程序。该工具箱允许您在 MATLAB 及其他工具箱中使用支持并行计算的函数。您还可以将其与 Simulink® 结合使用，以并行方式运行同一模型的多个仿真。程序和模型既可以在交互模式下运行，也可以以批处理模式运行。\n\n[XGBoost](https:\u002F\u002Fxgboost.readthedocs.io\u002F) 是一个优化的分布式梯度提升库，设计高效、灵活且易于移植。它在梯度提升框架下实现了多种机器学习算法。XGBoost 提供并行树提升（也称为 GBDT 或 GBM），能够快速而准确地解决许多数据科学问题。它支持在多台机器上进行分布式训练，包括 AWS、GCE、Azure 和 Yarn 集群。此外，它还可以与 Flink、Spark 等云数据流系统集成。\n\n[LIBSVM](https:\u002F\u002Fwww.csie.ntu.edu.tw\u002F~cjlin\u002Flibsvm\u002F) 是一个用于支持向量分类（C-SVC、nu-SVC）、回归（epsilon-SVR、nu-SVR）和分布估计（一类 SVM）的综合软件。它支持多类分类。\n\n[Azure Databricks](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fdatabricks\u002F) 是一种基于 Apache Spark 的快速协作式大数据分析服务，专为数据科学和数据工程设计。Azure Databricks 可以在几分钟内搭建 Apache Spark 环境，自动扩展规模，并在交互式工作区中协同处理共享项目。Azure Databricks 支持 Python、Scala、R、Java 和 SQL，以及 TensorFlow、PyTorch 和 scikit-learn 等数据科学框架和库。\n\n[Microsoft Cognitive Toolkit (CNTK)](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fcognitive-toolkit\u002F) 是一个面向商业级分布式深度学习的开源工具包。它通过有向图将神经网络描述为一系列计算步骤。CNTK 允许用户轻松实现和组合流行的模型类型，如前馈 DNN、卷积神经网络（CNN）和循环神经网络（RNN\u002FLSTM）。CNTK 实现了随机梯度下降（SGD，误差反向传播）学习，并支持自动微分及跨多个 GPU 和服务器的并行化。\n\n[Tensorflow_macOS](https:\u002F\u002Fgithub.com\u002Fapple\u002Ftensorflow_macos) 是针对 macOS 11.0 及以上版本优化的 TensorFlow 及 TensorFlow Addons 版本，利用 Apple 的 ML Compute 框架加速运行。\n\n[Apache Airflow](https:\u002F\u002Fairflow.apache.org) 是一个由社区创建的开源工作流管理平台，用于以编程方式编写、调度和监控工作流。安装简便、原则清晰、可扩展性强。Airflow 具有模块化架构，使用消息队列来编排任意数量的工作节点。Airflow 能够无限扩展。\n\n[Open Neural Network Exchange(ONNX)](https:\u002F\u002Fgithub.com\u002Fonnx) 是一个开放生态系统，使 AI 开发者能够根据项目的发展选择合适的工具。ONNX 提供了一种适用于深度学习和传统机器学习的开源模型格式，定义了一个可扩展的计算图模型，以及内置算子和标准数据类型的规范。\n\n[Apache MXNet](https:\u002F\u002Fmxnet.apache.org\u002F) 是一个兼顾效率和灵活性的深度学习框架。它允许混合符号式和命令式编程，以最大化效率和生产力。MXNet 的核心是一个动态依赖调度器，能够实时自动并行化符号式和命令式操作。在其之上还有一层图优化层，使符号式执行既快速又节省内存。MXNet 具有良好的可移植性和轻量化特性，能够有效扩展到多 GPU 和多台机器。支持 Python、R、Julia、Scala、Go、JavaScript 等多种语言。\n\n[AutoGluon](https:\u002F\u002Fautogluon.mxnet.io\u002Findex.html) 是一个自动化机器学习工具包，能够简化深度学习任务，帮助用户在应用中轻松获得强大的预测性能。只需几行代码，即可在表格数据、图像和文本数据上训练并部署高精度的深度学习模型。\n\n[Anaconda](https:\u002F\u002Fwww.anaconda.com\u002F) 是一个非常流行的数据科学平台，适用于机器学习和深度学习，使用户能够开发、训练和部署模型。\n\n[PlaidML](https:\u002F\u002Fgithub.com\u002Fplaidml\u002Fplaidml) 是一个先进且可移植的张量编译器，旨在让笔记本电脑、嵌入式设备或其他计算硬件支持不足或软件栈存在不友好许可限制的设备也能运行深度学习。\n\n[OpenCV](https:\u002F\u002Fopencv.org) 是一个高度优化的库，专注于实时计算机视觉应用。其 C++、Python 和 Java 接口支持 Linux、MacOS、Windows、iOS 和 Android。\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) 是一个基于 SciPy、NumPy 和 matplotlib 构建的 Python 机器学习模块，使得许多流行的机器学习算法的稳健且简单的实现更加容易。\n\n[Weka](https:\u002F\u002Fwww.cs.waikato.ac.nz\u002Fml\u002Fweka\u002F) 是一款开源机器学习软件，可通过图形用户界面、标准终端应用程序或 Java API 访问。它广泛应用于教学、科研和工业领域，内置了大量用于标准机器学习任务的工具，并且还能透明地访问 scikit-learn、R 和 Deeplearning4j 等知名工具箱。\n\n[Caffe](https:\u002F\u002Fgithub.com\u002FBVLC\u002Fcaffe) 是一个以表达性、速度和模块化为核心设计的深度学习框架。它由伯克利人工智能研究实验室（BAIR）、伯克利视觉与学习中心（BVLC）以及社区贡献者共同开发。\n\n[Theano](https:\u002F\u002Fgithub.com\u002FTheano\u002FTheano) 是一个 Python 库，允许高效地定义、优化和评估涉及多维数组的数学表达式，并与 NumPy 紧密集成。\n\n[Microsoft Project Bonsai](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fproject-bonsai\u002F) 是一个低代码 AI 平台，用于加速 AI 驱动的自动化开发，是 Microsoft 自主导航系统套件的一部分。Bonsai 用于构建能够提供操作指导或自主决策的 AI 组件，以优化工艺参数、提高生产效率并减少停机时间。\n\n[Microsoft AirSim](https:\u002F\u002Fmicrosoft.github.io\u002FAirSim\u002Flidar.html) 是一个基于 Unreal Engine 构建的无人机、汽车等模拟器（同时也有实验性的 Unity 版本）。AirSim 是开源、跨平台的，支持与 PX4 和 ArduPilot 等主流飞行控制器进行“软件在环”仿真，以及与 PX4 进行“硬件在环”仿真，从而实现物理和视觉上高度逼真的模拟。它被开发为一个 Unreal 插件，可以轻松集成到任何 Unreal 场景中。AirSim 正在发展成为一个 AI 研究平台，用于试验自动驾驶车辆相关的深度学习、计算机视觉和强化学习算法。\n\n[CARLA](https:\u002F\u002Fgithub.com\u002Fcarla-simulator\u002Fcarla) 是一款用于自动驾驶研究的开源模拟器。CARLA 从零开始开发，旨在支持自动驾驶系统的开发、训练和验证。除了开源代码和通信协议外，CARLA 还提供了专为此目的创建的开放数字资产（城市布局、建筑物、车辆），用户可以免费使用。\n\n[CARLA 的 ROS\u002FROS2 桥接包](https:\u002F\u002Fgithub.com\u002Fcarla-simulator\u002Fros-bridge) 是一个实现 ROS 与 CARLA 之间双向通信的桥梁。CARLA 服务器中的信息会被转换为 ROS 主题；同样地，ROS 节点之间发送的消息也会被转换为可在 CARLA 中执行的命令。\n\n[ROS 工具箱](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fros.html) 是一种将 MATLAB® 和 Simulink® 与机器人操作系统（ROS 和 ROS 2）连接起来的工具，允许用户构建 ROS 节点网络。该工具箱包含 MATLAB 函数和 Simulink 模块，用于导入、分析和回放以 rosbag 文件格式记录的 ROS 数据。此外，还可以连接到实时 ROS 网络以访问 ROS 消息。\n\n[机器人技术工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Frobotics.html) 提供了一套专门针对机器人技术的功能（设计、仿真和测试机械臂、移动机器人和人形机器人），充分利用了 MATLAB 的原生能力（线性代数、跨平台兼容性和图形功能）。该工具箱还为移动机器人提供了多种功能，包括自行车运动模型、路径规划算法（bug 算法、距离变换、D*、PRM）、动力学规划（格网搜索、RRT）、定位（EKF、粒子滤波器）、地图构建（EKF）以及同时定位与地图构建（EKF），并包含一个非完整约束车辆的 Simulink 模型。此外，该工具箱还提供了一个详细的四旋翼飞行机器人的 Simulink 模型。\n\n[图像处理工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fimage.html) 是一种工具，提供了一整套符合行业标准的算法和工作流应用程序，用于图像处理、分析、可视化及算法开发。用户可以进行图像分割、图像增强、降噪、几何变换、图像配准以及三维图像处理等操作。\n\n[计算机视觉工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fcomputer-vision.html) 提供用于设计和测试计算机视觉、三维视觉和视频处理系统的算法、函数和应用程序。用户可以执行目标检测与跟踪，以及特征检测、提取和匹配。此外，还可以自动化单目、双目和鱼眼相机的标定流程。对于三维视觉，该工具箱支持视觉 SLAM 和点云 SLAM、立体视觉、运动恢复结构以及点云处理等功能。\n\n[机器人技术工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Frobotics.html) 是一种工具，它将机器人技术领域的特定功能（设计、仿真和测试机械臂、移动机器人和人形机器人）引入 MATLAB，并充分利用 MATLAB 的原生能力（线性代数、跨平台兼容性和图形功能）。该工具箱还为移动机器人提供了自行车运动模型、路径规划算法（bug 算法、距离变换、D*、PRM）、动力学规划（格网搜索、RRT）、定位（EKF、粒子滤波器）、地图构建（EKF）以及同时定位与地图构建（EKF）等功能，并包含一个非完整约束车辆的 Simulink 模型。此外，该工具箱还提供了一个详细的四旋翼飞行机器人的 Simulink 模型。\n\n[模型预测控制工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmodel-predictive-control.html) 提供用于设计和仿真线性和非线性模型预测控制器（MPC）的函数、应用程序和 Simulink® 模块。该工具箱允许用户指定被控对象和扰动模型、预测时域、约束条件以及权重。通过运行闭环仿真，用户可以评估控制器的性能。\n\n[预测性维护工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fpredictive-maintenance.html) 允许用户管理传感器数据、设计状态指标，并估算设备的剩余使用寿命（RUL）。该工具箱提供基于数据和基于模型的技术函数及交互式应用程序，用于探索、提取和排序特征，包括统计分析、频谱分析和时间序列分析等方法。\n\n[Vision HDL 工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fvision-hdl.html) 是一种工具，提供用于在 FPGA 和 ASIC 上设计和实现视觉系统的像素流算法。它提供一个设计框架，支持多种接口类型、帧尺寸和帧率。该工具箱中的图像处理、视频和计算机视觉算法采用了适合 HDL 实现的架构。\n\n[自动驾驶工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fautomated-driving.html) 是一种 MATLAB 工具，提供用于设计、仿真和测试 ADAS 及自动驾驶系统的算法和工具。用户可以设计和测试视觉和激光雷达感知系统，以及传感器融合、路径规划和车辆控制器。可视化工具包括鸟瞰图、传感器覆盖范围、检测结果和轨迹的示意图，以及用于显示视频、激光雷达数据和地图的界面。该工具箱支持导入和使用 HERE HD Live Map 数据和 OpenDRIVE® 道路网络。此外，它还提供了常见 ADAS 和自动驾驶功能的参考应用示例，包括前向碰撞预警（FCW）、自动紧急制动（AEB）、自适应巡航控制（ACC）、车道保持辅助（LKA）以及代客泊车等功能。该工具箱支持 C\u002FC++ 代码生成，用于快速原型开发和 HIL 测试，并且对传感器融合、跟踪、路径规划和车辆控制器算法提供支持。\n\n[导航工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fnavigation.html) 提供用于运动规划、同时定位与地图构建（SLAM）以及惯性导航的算法和分析工具。该工具箱包含可定制的基于搜索和采样的路径规划器，以及用于验证和比较路径的度量标准。用户可以创建二维和三维地图表示，利用 SLAM 算法生成地图，并通过 SLAM 地图构建应用程序交互式地可视化和调试地图生成过程。\n\n[UAV 工具箱](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fuav.html) 是一款应用程序，提供用于设计、仿真、测试和部署无人机及无人飞行器应用的工具和参考应用。用户可以设计自主飞行算法、无人机任务和飞行控制器。飞行日志分析器应用程序允许用户交互式地分析常见飞行日志格式中的三维飞行轨迹、遥测信息和传感器读数。\n\n[Lidar 工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Flidar.html) 是一款提供算法、函数和应用程序的工具，用于设计、分析和测试激光雷达处理系统。您可以执行目标检测与跟踪、语义分割、形状拟合、激光雷达配准以及障碍物检测。Lidar 工具箱支持激光雷达与相机的交叉标定，适用于结合计算机视觉和激光雷达处理的工作流程。\n\n[Mapping 工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmapping.html) 是一款提供地理数据转换算法和函数、并用于创建地图显示的工具。您可以在地理背景下可视化数据，基于超过60种地图投影构建地图显示，并将来自各种来源的数据转换为一致的地理坐标系。\n\n\n\n\n# 计算机视觉开发\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_01c04b532852.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## 计算机视觉学习资源\n\n[计算机视觉](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Foverview\u002Fwhat-is-computer-vision\u002F) 是人工智能（AI）的一个领域，专注于使计算机能够识别和理解图像及视频中的物体和人物。\n\n[OpenCV 课程](https:\u002F\u002Fopencv.org\u002Fcourses\u002F)\n\n[在 Microsoft Azure 中探索计算机视觉](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Flearn\u002Fpaths\u002Fexplore-computer-vision-microsoft-azure\u002F)\n\n[在线顶级计算机视觉课程 | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?languages=en&query=computer%20vision)\n\n[在线顶级计算机视觉课程 | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fcomputer-vision\u002F)\n\n[通过在线课程和教程学习计算机视觉 | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fcomputer-vision)\n\n[计算机视觉与图像处理基础 | edX](https:\u002F\u002Fwww.edx.org\u002Fcourse\u002Fcomputer-vision-and-image-processing-fundamentals)\n\n[计算机视觉入门课程 | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fintroduction-to-computer-vision--ud810)\n\n[计算机视觉纳米学位项目 | Udacity](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fcomputer-vision-nanodegree--nd891)\n\n[机器视觉课程 | MIT 开放式课程](https:\u002F\u002Focw.mit.edu\u002Fcourses\u002Felectrical-engineering-and-computer-science\u002F6-801-machine-vision-fall-2004\u002F)\n\n[计算机视觉培训课程 | NobleProg](https:\u002F\u002Fwww.nobleprog.com\u002Fcomputer-vision-training)\n\n[视觉计算研究生项目 | 斯坦福在线](https:\u002F\u002Fonline.stanford.edu\u002Fprograms\u002Fvisual-computing-graduate-program)\n\n## 计算机视觉工具、库和框架\n\n[OpenCV](https:\u002F\u002Fopencv.org) 是一个高度优化的库，专注于实时计算机视觉应用。其 C++、Python 和 Java 接口支持 Linux、MacOS、Windows、iOS 和 Android。\n\n[Microsoft 计算机视觉示例](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fcomputervision-recipes) 是一个项目，提供了构建计算机视觉系统的示例和最佳实践指南。该项目旨在打造一套全面的工具和示例，利用计算机视觉算法、神经网络架构以及系统部署方面的最新进展，基于现有的先进库进行开发，并围绕图像数据加载、模型优化与评估以及向云端扩展等功能构建附加工具。\n\n[Microsoft 认知工具包（CNTK）](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fcognitive-toolkit\u002F) 是一个面向商业级分布式深度学习的开源工具包。它通过有向图将神经网络描述为一系列计算步骤。CNTK 允许用户轻松实现和组合流行的模型类型，如前馈 DNN、卷积神经网络（CNN）和循环神经网络（RNN\u002FLSTM）。CNTK 实现了随机梯度下降（SGD，误差反向传播）学习，并具备自动微分功能，可在多个 GPU 和服务器上并行化。\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) 是一个基于 SciPy、NumPy 和 matplotlib 构建的 Python 机器学习模块，使得许多流行机器学习算法的稳健且简单的实现更加容易。\n\n[NVIDIA cuDNN](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcudnn) 是一个针对 [深度神经网络](https:\u002F\u002Fdeveloper.nvidia.com\u002Fdeep-learning) 的 GPU 加速原语库。cuDNN 提供了对标准操作的高度优化实现，例如前向和反向卷积、池化、归一化和激活层。cuDNN 可加速广泛使用的深度学习框架，包括 [Caffe2](https:\u002F\u002Fcaffe2.ai\u002F)、[Chainer](https:\u002F\u002Fchainer.org\u002F)、[Keras](https:\u002F\u002Fkeras.io\u002F)、[MATLAB](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fdeep-learning.html)、[MxNet](https:\u002F\u002Fmxnet.incubator.apache.org\u002F)、[PyTorch](https:\u002F\u002Fpytorch.org\u002F) 和 [TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002F) 等。\n\n[自动驾驶工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fautomated-driving.html) 是 MATLAB 提供的一款工具，用于设计、仿真和测试 ADAS 及自动驾驶系统。您可以设计和测试视觉与激光雷达感知系统，以及传感器融合、路径规划和车辆控制器。可视化工具包括鸟瞰图、传感器覆盖范围、检测结果与跟踪轨迹等视图，以及用于显示视频、激光雷达和地图的界面。该工具箱允许导入并使用 HERE HD Live Map 数据和 OpenDRIVE® 道路网络。它还提供了常见 ADAS 和自动驾驶功能的参考应用示例，包括 FCW、AEB、ACC、LKA 和代客泊车等。该工具箱支持 C\u002FC++ 代码生成，以实现快速原型制作和 HIL 测试，并支持传感器融合、跟踪、路径规划和车辆控制器算法。\n\n[LRSLibrary](https:\u002F\u002Fgithub.com\u002Fandrewssobral\u002Flrslibrary) 是一个用于视频中背景建模与减除的低秩稀疏工具库。该库最初设计用于视频中的运动目标检测，但也可用于其他计算机视觉和机器学习问题。\n\n[图像处理工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fimage.html) 是一款提供全面的参考标准算法和工作流应用程序的工具，用于图像处理、分析、可视化和算法开发。您可以执行图像分割、图像增强、降噪、几何变换、图像配准以及三维图像处理。\n\n[计算机视觉工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fcomputer-vision.html) 是一款提供算法、函数和应用程序的工具，用于设计和测试计算机视觉、三维视觉以及视频处理系统。您可以执行目标检测与跟踪、特征检测、提取及匹配等操作。此外，还可以自动化单目、双目和鱼眼相机的标定流程。对于三维视觉，该工具箱支持视觉与点云 SLAM、立体视觉、运动恢复结构以及点云处理。\n\n[统计与机器学习工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fstatistics.html) 是一款提供函数和应用程序的工具，用于描述、分析和建模数据。您可以通过描述性统计、可视化和聚类进行探索性数据分析；将概率分布拟合到数据；生成用于蒙特卡洛模拟的随机数，并执行假设检验。回归和分类算法使您能够从数据中得出推论，并通过交互式的方式（使用分类与回归学习器应用程序）或编程方式（使用 AutoML）构建预测模型。\n\n[LiDAR 工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Flidar.html) 是一款提供算法、函数和应用程序的工具，用于设计、分析和测试 LiDAR 处理系统。您可以执行目标检测与跟踪、语义分割、形状拟合、LiDAR 数据配准以及障碍物检测等任务。LiDAR 工具箱支持结合计算机视觉和 LiDAR 处理的工作流中的 LiDAR 与相机交叉标定。\n\n[地图工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmapping.html) 是一款提供算法和函数的工具，用于转换地理数据并创建地图显示。您可以在地理背景下可视化数据，基于超过 60 种地图投影构建地图显示，并将来自各种来源的数据转换为一致的地理坐标系。\n\n[UAV 工具箱](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fuav.html) 是一款提供工具和参考应用的软件，用于设计、仿真、测试和部署无人机（UAV）及无人飞行器应用。您可以设计自主飞行算法、无人机任务和飞行控制器。飞行日志分析器应用程序允许您以交互方式分析常见飞行日志格式中的三维飞行轨迹、遥测信息和传感器读数。\n\n[并行计算工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmatlab-parallel-server.html) 是一款使您能够利用多核处理器、GPU 和计算机集群解决计算密集型和数据密集型问题的工具。诸如并行 for 循环、特殊数组类型和并行化数值算法等高级构造，使您无需 CUDA 或 MPI 编程即可实现 MATLAB® 应用程序的并行化。该工具箱允许您在 MATLAB 及其他工具箱中使用支持并行计算的函数。您还可以将其与 Simulink® 结合使用，以并行运行同一模型的多个仿真。程序和模型既可在交互模式下运行，也可在批处理模式下运行。\n\n[偏微分方程工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fpde.html) 是一款提供函数的工具，用于通过有限元分析求解结构力学、传热以及一般偏微分方程（PDE）问题。\n\n[ROS 工具箱](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fros.html) 是一款提供接口的工具，可将 MATLAB® 和 Simulink® 与机器人操作系统（ROS 和 ROS 2）连接起来，从而帮助您创建 ROS 节点网络。该工具箱包含 MATLAB 函数和 Simulink 模块，可用于导入、分析和回放以 rosbag 文件格式记录的 ROS 数据。您还可以连接到实时 ROS 网络，以访问 ROS 消息。\n\n[机器人技术工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Frobotics.html) 提供了一套工具，将机器人技术相关的功能（如机械臂、移动机器人和人形机器人的设计、仿真和测试）引入 MATLAB，并充分利用 MATLAB 的原生能力（线性代数、可移植性和图形功能）。该工具箱还为移动机器人提供了机器人运动模型（自行车模型）、路径规划算法（bug 算法、距离变换、D* 算法、PRM）、动力学规划（格子规划、RRT）、定位（EKF、粒子滤波器）、地图构建（EKF）以及同时定位与地图构建（EKF）等功能，并包含一个非完整约束车辆的 Simulink 模型。此外，该工具箱还提供了一个详细的四旋翼飞行机器人 Simulink 模型。\n\n[深度学习工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdeep-learning.html) 是一款提供框架的工具，用于借助算法、预训练模型和应用程序设计并实现深度神经网络。您可以使用卷积神经网络（ConvNets、CNNs）和长短期记忆网络（LSTM）对图像、时间序列和文本数据进行分类和回归。借助自动微分、自定义训练循环和共享权重，您可以构建生成对抗网络（GANs）和暹罗网络等网络架构。通过深度网络设计器应用程序，您可以以图形化方式设计、分析和训练网络。该工具箱可通过 ONNX 格式与 TensorFlow™ 和 PyTorch 进行模型交换，并可导入来自 TensorFlow-Keras 和 Caffe 的模型。该工具箱支持迁移学习，兼容 DarkNet-53、ResNet-50、NASNet、SqueezeNet 等多种预训练模型。\n\n[强化学习工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Freinforcement-learning.html) 是一款提供应用程序、函数和 Simulink® 块的工具，用于使用强化学习算法（包括 DQN、PPO、SAC 和 DDPG）训练策略。您可以利用这些策略为资源分配、机器人技术和自主系统等复杂应用实现控制器和决策算法。\n\n[深度学习 HDL 工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdeep-learning-hdl.html) 是一款提供函数和工具的工具，用于在 FPGA 和 SoC 上原型化并实现深度学习网络。它提供了预构建的比特流，可在受支持的 Xilinx® 和 Intel® FPGA 以及 SoC 设备上运行多种深度学习网络。性能分析和估算工具使您能够通过权衡设计、性能和资源利用率之间的关系来定制深度学习网络。\n\n[模型预测控制工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmodel-predictive-control.html) 是一款提供函数、应用程序和 Simulink® 块的工具，用于使用线性和非线性模型预测控制（MPC）设计和仿真控制器。该工具箱允许您指定被控对象和干扰模型、预测时域、约束条件以及权重。通过运行闭环仿真，您可以评估控制器的性能。\n\n[Vision HDL Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fvision-hdl.html) 是一款工具，为在 FPGA 和 ASIC 上设计与实现视觉系统提供像素流算法。它提供了一个设计框架，支持多种接口类型、帧尺寸和帧率。该工具箱中的图像处理、视频和计算机视觉算法采用适合 HDL 实现的架构。\n\n[Data Acquisition Toolbox™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdata-acquisition.html) 是一款工具，提供用于配置数据采集硬件、将数据读入 MATLAB® 和 Simulink® 以及将数据写入 DAQ 模拟和数字输出通道的应用程序和函数。该工具箱支持多种 DAQ 硬件，包括来自 National Instruments® 及其他供应商的 USB、PCI、PCI Express®、PXI® 和 PXI Express® 设备。\n\n[Microsoft AirSim](https:\u002F\u002Fmicrosoft.github.io\u002FAirSim\u002Flidar.html) 是一款基于虚幻引擎（并有实验性的 Unity 版本）开发的无人机、汽车等模拟器。AirSim 是开源、跨平台的，支持与流行的飞行控制器（如 PX4 和 ArduPilot）进行软件在环仿真，以及与 PX4 进行硬件在环仿真，从而实现物理和视觉上高度逼真的模拟。它被开发为一个虚幻插件，可以轻松集成到任何虚幻环境中。AirSim 正在被打造成为一个 AI 研究平台，用于试验深度学习、计算机视觉和强化学习算法在自动驾驶车辆中的应用。\n\n\n\n# 自然语言处理开发\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_c20b531e32f3.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## 自然语言处理学习资源\n\n[自然语言处理 (NLP)](https:\u002F\u002Fwww.ibm.com\u002Fcloud\u002Flearn\u002Fnatural-language-processing) 是人工智能 (AI) 的一个分支，专注于赋予计算机像人类一样理解文本和口语的能力。NLP 将基于规则的人类语言计算语言学建模与统计、机器学习和深度学习模型相结合。\n\n[使用 Python 的 NLTK 包进行自然语言处理](https:\u002F\u002Frealpython.com\u002Fnltk-nlp-python\u002F)\n\n[Cognitive Services—面向 AI 开发者的 API | Microsoft Azure](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fcognitive-services\u002F)\n\n[人工智能服务 - 亚马逊网络服务 (AWS)](https:\u002F\u002Faws.amazon.com\u002Fmachine-learning\u002Fai-services\u002F)\n\n[Google Cloud 自然语言处理 API](https:\u002F\u002Fcloud.google.com\u002Fnatural-language\u002Fdocs\u002Freference\u002Frest)\n\n[在线顶级自然语言处理课程 | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fnatural-language-processing\u002F)\n\n[自然语言处理 (NLP) 入门 | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fnatural-language-processing\u002F)\n\n[顶级自然语言处理课程 | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fcourses?=&query=natural%20language%20processing)\n\n[自然语言处理 | Coursera](https:\u002F\u002Fwww.coursera.org\u002Flearn\u002Flanguage-processing)\n\n[TensorFlow 中的自然语言处理 | Coursera](https:\u002F\u002Fwww.coursera.org\u002Flearn\u002Fnatural-language-processing-tensorflow)\n\n[通过在线课程和教程学习自然语言处理 | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fnatural-language-processing)\n\n[使用 Microsoft Azure 构建自然语言处理解决方案 | Pluralsight](https:\u002F\u002Fwww.pluralsight.com\u002Fcourses\u002Fbuild-natural-language-processing-solution-microsoft-azure)\n\n[自然语言处理 (NLP) 培训课程 | NobleProg](https:\u002F\u002Fwww.nobleprog.com\u002Fnlp-training)\n\n[斯坦福在线：深度学习与自然语言处理课程](https:\u002F\u002Fonline.stanford.edu\u002Fcourses\u002Fcs224n-natural-language-processing-deep-learning)\n\n[麻省理工学院 OpenCourseWare：高级自然语言处理](https:\u002F\u002Focw.mit.edu\u002Fcourses\u002Felectrical-engineering-and-computer-science\u002F6-864-advanced-natural-language-processing-fall-2005\u002F)\n\n[认证自然语言处理专家证书 | IABAC](https:\u002F\u002Fiabac.org\u002Fartificial-intelligence-certification\u002Fcertified-natural-language-processing-expert\u002F)\n\n[英特尔自然语言处理课程](https:\u002F\u002Fsoftware.intel.com\u002Fcontent\u002Fwww\u002Fus\u002Fen\u002Fdevelop\u002Ftraining\u002Fcourse-natural-language-processing.html)\n\n\n## 自然语言处理工具、库和框架\n\n[Natural Language Toolkit (NLTK)](https:\u002F\u002Fwww.nltk.org\u002F) 是一个用于构建处理人类语言数据的 Python 程序的领先平台。它提供了易于使用的接口，可访问超过 [50 个语料库和词汇资源](https:\u002F\u002Fnltk.org\u002Fnltk_data\u002F)，例如 WordNet，并附带一套用于分类、分词、词干提取、词性标注、句法分析和语义推理的文本处理库，以及工业级 NLP 库的封装器。\n\n[spaCy](https:\u002F\u002Fspacy.io) 是一个用于 Python 和 Cython 的高级自然语言处理库。它基于最新的研究成果构建，从一开始就被设计用于实际产品中。spaCy 配备了预训练好的管道，目前支持 60 多种语言的分词和训练。它还具有用于词性标注、句法分析、命名实体识别、文本分类等任务的神经网络模型，并支持与 BERT 等预训练转换器进行多任务学习。\n\n[CoreNLP](https:\u002F\u002Fstanfordnlp.github.io\u002FCoreNLP\u002F) 是一组用 Java 编写的自然语言分析工具。CoreNLP 使用户能够为文本生成语言学注释，包括词和句子边界、词性、命名实体、数值和时间值、依存关系和结构关系解析、指代消解、情感、引文归属及各种关系。\n\n[NLPnet](https:\u002F\u002Fgithub.com\u002Ferickrf\u002Fnlpnet) 是一个基于神经网络的 Python 自然语言处理库。它可以执行词性标注、语义角色标注和依存关系解析。\n\n[Flair](https:\u002F\u002Fgithub.com\u002FflairNLP\u002Fflair) 是一个简单的框架，可用于将最先进的自然语言处理 (NLP) 模型应用于您的文本，例如命名实体识别 (NER)、词性标注 (PoS)、生物医学数据的特殊支持、语义消歧义和分类，并且支持的语言数量正在迅速增加。\n\n[Catalyst](https:\u002F\u002Fgithub.com\u002Fcuriosity-ai\u002Fcatalyst) 是一个以速度为导向的 C# 自然语言处理库。它受到 [spaCy 的设计](https:\u002F\u002Fspacy.io\u002F) 启发，带来了预训练模型、开箱即用的词嵌入和文档嵌入训练支持，以及灵活的实体识别模型。\n\n[Apache OpenNLP](https:\u002F\u002Fopennlp.apache.org\u002F) 是一个基于机器学习的自然语言文本处理开源工具库。它提供了一系列 API，可用于命名实体识别、句子检测、词性标注、分词、特征提取、组块分析、句法分析以及指代消解等任务。\n\n[Microsoft Cognitive Toolkit (CNTK)](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fcognitive-toolkit\u002F) 是一个面向商业级应用的开源分布式深度学习工具包。它通过有向图将神经网络表示为一系列计算步骤。CNTK 允许用户轻松实现并组合常用的模型类型，如前馈深度神经网络、卷积神经网络（CNN）和循环神经网络（RNN\u002FLSTM）。CNTK 实现了带有自动微分功能的随机梯度下降（SGD，即误差反向传播）训练，并支持在多 GPU 和多台服务器上的并行化。\n\n[NVIDIA cuDNN](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcudnn) 是一个针对 [深度神经网络](https:\u002F\u002Fdeveloper.nvidia.com\u002Fdeep-learning) 的 GPU 加速原语库。cuDNN 提供了对标准操作的高度优化实现，例如前向和反向卷积、池化、归一化以及激活层。cuDNN 可以加速广泛使用的深度学习框架，包括 [Caffe2](https:\u002F\u002Fcaffe2.ai\u002F)、[Chainer](https:\u002F\u002Fchainer.org\u002F)、[Keras](https:\u002F\u002Fkeras.io\u002F)、[MATLAB](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fdeep-learning.html)、[MxNet](https:\u002F\u002Fmxnet.incubator.apache.org\u002F)、[PyTorch](https:\u002F\u002Fpytorch.org\u002F) 和 [TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002F) 等。\n\n[TensorFlow](https:\u002F\u002Fwww.tensorflow.org) 是一个端到端的开源机器学习平台。它拥有全面且灵活的工具、库和社区资源生态系统，使研究人员能够推动机器学习领域的前沿发展，同时也让开发者能够轻松构建和部署基于机器学习的应用程序。\n\n[Tensorflow_macOS](https:\u002F\u002Fgithub.com\u002Fapple\u002Ftensorflow_macos) 是 TensorFlow 及其扩展库的 macOS 优化版本，适用于 macOS 11.0 及以上系统，并利用 Apple 的 ML Compute 框架进行加速。\n\n[Keras](https:\u002F\u002Fkeras.io) 是一个用 Python 编写的高级神经网络 API，可运行在 TensorFlow、CNTK 或 Theano 等后端之上。它专注于支持快速实验，能够在 TensorFlow、Microsoft Cognitive Toolkit、R、Theano 或 PlaidML 等平台上运行。\n\n[PyTorch](https:\u002F\u002Fpytorch.org) 是一个用于处理不规则输入数据（如图、点云和流形）的深度学习库，主要由 Facebook 的 AI 研究实验室开发。\n\n[Eclipse Deeplearning4J (DL4J)](https:\u002F\u002Fdeeplearning4j.konduit.ai\u002F) 是一组旨在支持基于 JVM（Scala、Kotlin、Clojure 和 Groovy）的深度学习应用所有需求的项目。这意味着从原始数据开始，无论数据来自何处、采用何种格式，都可以对其进行加载和预处理，进而构建和调优各种简单或复杂的深度学习网络。\n\n[Chainer](https:\u002F\u002Fchainer.org\u002F) 是一个以灵活性为目标的 Python 深度学习框架。它提供了基于“定义即运行”方法（动态计算图）的自动微分 API，以及面向对象的高级 API 来构建和训练神经网络。此外，它还通过 [CuPy](https:\u002F\u002Fgithub.com\u002Fcupy\u002Fcupy) 支持 CUDA\u002FcuDNN，以实现高性能的训练和推理。\n\n[Anaconda](https:\u002F\u002Fwww.anaconda.com\u002F) 是一个非常流行的数据科学平台，专为机器学习和深度学习设计，帮助用户开发、训练和部署模型。\n\n[PlaidML](https:\u002F\u002Fgithub.com\u002Fplaidml\u002Fplaidml) 是一种先进且可移植的张量编译器，能够在笔记本电脑、嵌入式设备或其他计算硬件支持不足或软件许可限制较多的设备上运行深度学习。\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) 是一个基于 SciPy、NumPy 和 matplotlib 构建的 Python 机器学习模块，使得应用许多流行的机器学习算法的稳健且简单的实现变得更加容易。\n\n[Caffe](https:\u002F\u002Fgithub.com\u002FBVLC\u002Fcaffe) 是一个以表达性、速度和模块化为核心设计的深度学习框架。它由伯克利人工智能研究实验室（BAIR）\u002F伯克利视觉与学习中心（BVLC）及社区贡献者共同开发。\n\n[Theano](https:\u002F\u002Fgithub.com\u002FTheano\u002FTheano) 是一个 Python 库，允许高效地定义、优化和评估涉及多维数组的数学表达式，并与 NumPy 紧密集成。\n\n[Apache Spark](https:\u002F\u002Fspark.apache.org\u002F) 是一个用于大规模数据处理的统一分析引擎。它提供了 Scala、Java、Python 和 R 四种高级 API，以及一个优化的执行引擎，支持通用计算图进行数据分析。此外，它还支持丰富的高层工具，包括用于 SQL 和 DataFrame 的 Spark SQL、用于机器学习的 MLlib、用于图处理的 GraphX 以及用于流式处理的 Structured Streaming。\n\n[Apache Spark Connector for SQL Server and Azure SQL](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fsql-spark-connector) 是一个高性能连接器，使用户能够在大数据分析中使用事务型数据，并将结果持久化以供即席查询或报告使用。该连接器允许将任何 SQL 数据库（无论是本地部署还是云端）用作 Spark 作业的输入数据源或输出数据接收端。\n\n[Apache PredictionIO](https:\u002F\u002Fpredictionio.apache.org\u002F) 是一个面向开发者、数据科学家和最终用户的开源机器学习框架。它支持事件收集、算法部署、评估以及通过 REST API 查询预测结果等功能。该框架基于 Hadoop、HBase（以及其他数据库）、Elasticsearch、Spark 等可扩展的开源服务，并实现了所谓的 Lambda 架构。\n\n[Apache Airflow](https:\u002F\u002Fairflow.apache.org) 是一个由社区创建的开源工作流管理平台，用于以编程方式编写、调度和监控工作流。Airflow 具有模块化架构，使用消息队列来协调任意数量的工作节点。它能够无限扩展。\n\n[Open Neural Network Exchange (ONNX)](https:\u002F\u002Fgithub.com\u002Fonnx) 是一个开放的生态系统，使 AI 开发人员能够根据项目的发展阶段选择合适的工具。ONNX 提供了一种开源的 AI 模型格式，适用于深度学习和传统机器学习。它定义了一个可扩展的计算图模型，以及内置算子和标准数据类型的规范。\n\n[BigDL](https:\u002F\u002Fbigdl-project.github.io\u002F) 是一个面向 Apache Spark 的分布式深度学习库。借助 BigDL，用户可以将深度学习应用编写为标准的 Spark 程序，这些程序可以直接在现有的 Spark 或 Hadoop 集群上运行。\n\n[Numba](https:\u002F\u002Fgithub.com\u002Fnumba\u002Fnumba) 是由 Anaconda, Inc. 赞助的一个开源、支持 NumPy 的 Python 优化编译器。它利用 LLVM 编译器项目，将 Python 语法转换为机器码。Numba 可以编译大量以数值计算为主的 Python 代码，包括许多 NumPy 函数。此外，Numba 还支持循环的自动并行化、生成 GPU 加速代码，以及创建 ufunc 和 C 回调函数。\n\n\n\n# 生物信息学\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_cd7bb1386d0d.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## 生物信息学学习资源\n\n[生物信息学](https:\u002F\u002Fwww.genome.gov\u002Fgenetics-glossary\u002FBioinformatics) 是一门与生物分子序列分析相关的计算科学领域。这通常涉及基因、DNA、RNA 或蛋白质等分子，尤其适用于比较同一生物体内部或不同生物体之间的基因及其他序列，研究生物间的进化关系，并通过 DNA 和蛋白质序列中的模式来推断其功能。\n\n[欧洲生物信息学研究所](https:\u002F\u002Fwww.ebi.ac.uk\u002F)\n\n[美国国家生物技术信息中心](https:\u002F\u002Fwww.ncbi.nlm.nih.gov)\n\n[生物信息学在线课程 | ISCB - 国际计算生物学学会](https:\u002F\u002Fwww.iscb.org\u002Fcms_addon\u002Fonline_courses\u002Findex.php)\n\n[生物信息学 | Coursera](https:\u002F\u002Fwww.coursera.org\u002Fspecializations\u002Fbioinformatics)\n\n[顶级生物信息学课程 | Udemy](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002FBioinformatics\u002F)\n\n[生物识别技术课程 | Udemy](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fbiometrics\u002F)\n\n[通过在线课程和教程学习生物信息学 | edX](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fbioinformatics)\n\n[哈佛大学扩展学院生物信息学研究生证书](https:\u002F\u002Fextension.harvard.edu\u002Facademics\u002Fprograms\u002Fbioinformatics-graduate-certificate\u002F)\n\n[加州大学圣地亚哥分校扩展学院：生物信息学与生物统计学](https:\u002F\u002Fextension.ucsd.edu\u002Fcourses-and-programs\u002Fbioinformatics-and-biostatistics)\n\n[生物信息学与蛋白质组学——免费在线课程资料 | MIT](https:\u002F\u002Focw.mit.edu\u002Fcourses\u002Felectrical-engineering-and-computer-science\u002F6-092-bioinformatics-and-proteomics-january-iap-2005\u002F)\n\n[生物识别技术入门课程 — 生物识别研究所](https:\u002F\u002Fwww.biometricsinstitute.org\u002Fevent\u002Fintroduction-to-biometrics-short-course\u002F)\n\n## 生物信息学工具、库和框架\n\n[Bioconductor](https:\u002F\u002Fbioconductor.org\u002F) 是一个开源项目，提供用于分析和理解高通量基因组数据的工具。Bioconductor 使用 [R 统计编程语言](https:\u002F\u002Fwww.r-project.org\u002Fabout.html)，采用开源和开放开发模式。它每年发布两次版本，并拥有活跃的用户社区。Bioconductor 也以 [AMI（Amazon Machine Image）](https:\u002F\u002Fdocs.aws.amazon.com\u002FAWSEC2\u002Flatest\u002FUserGuide\u002FAMIs.html) 和 [Docker 镜像](https:\u002F\u002Fdocs.docker.com\u002Fengine\u002Freference\u002Fcommandline\u002Fimages\u002F) 的形式提供。\n\n[Bioconda](https:\u002F\u002Fbioconda.github.io) 是 conda 软件包管理器的一个频道，专门用于生物信息学软件。它拥有一个包含超过 7000 个生物信息学软件包的仓库，可以直接通过 conda install 命令安装使用。\n\n[UniProt](https:\u002F\u002Fwww.uniprot.org\u002F) 是一个免费开放的数据库，为用户提供全面、高质量且可自由访问的蛋白质序列集合，并附有功能注释信息。\n\n[Bowtie 2](https:\u002F\u002Fbio.tools\u002Fbowtie2#!) 是一款超快速且内存高效的比对工具，用于将测序读段比对到长参考序列上。它特别擅长比对长度约为 50 到数百或数千碱基的读段，尤其适用于比对相对较长的（哺乳动物）基因组。\n\n[Biopython](https:\u002F\u002Fbiopython.org\u002F) 是由国际开发者团队用 Python 编写的、可用于生物计算的一系列免费工具。它是一个分布式协作项目，旨在开发满足当前和未来生物信息学需求的 Python 库和应用程序。\n\n[BioRuby](https:\u002F\u002Fbioruby.open-bio.org\u002F) 是一个工具包，包含用于序列分析、通路分析、蛋白质建模和系统发育分析的组件；它支持多种广泛使用的数据格式，并能方便地访问数据库、外部程序和公共网络服务，包括 BLAST、KEGG、GenBank、MEDLINE 和 GO 等。\n\n[BioJava](https:\u002F\u002Fbiojava.org\u002F) 是一个工具包，提供 API 来维护本地 PDB 数据库的安装、加载和操作结构、执行序列和结构比对等标准分析，并以 3D 方式进行可视化。\n\n[BioPHP](https:\u002F\u002Fbiophp.org\u002F) 是一个开源项目，提供一系列开源 PHP 代码，包含用于 DNA 和蛋白质序列分析、比对、数据库解析以及其他生物信息学工具的类。\n\n[Avogadro](https:\u002F\u002Favogadro.cc\u002F) 是一款先进的分子编辑器和可视化工具，专为跨平台使用而设计，适用于计算化学、分子建模、生物信息学、材料科学及相关领域。它提供灵活的高质量渲染和强大的插件架构。\n\n[Ascalaph Designer](https:\u002F\u002Fwww.biomolecular-modeling.com\u002FAscalaph\u002FAscalaph_Designer.html) 是一款用于分子动力学模拟的程序。在单一图形界面下，既实现了自己的分子动力学算法，也集成了主流程序中的经典力学和量子力学方法。\n\n[Anduril](https:\u002F\u002Fwww.anduril.org\u002Fsite\u002F) 是一个用于分析大型数据集的工作流平台。Anduril 提供了在生物医学研究中分析高通量数据的功能，并且该平台可以被第三方完全扩展。现成的工具支持数据可视化、DNA\u002FRNA\u002FChIP 测序、DNA\u002FRNA 微阵列、细胞计数和图像分析等任务。\n\n[Galaxy](https:\u002F\u002Fmelbournebioinformatics.github.io\u002FMelBioInf_docs\u002Ftutorials\u002Fgalaxy_101\u002Fgalaxy_101\u002F) 是一个开源的基于 Web 的平台，用于开展可访问、可重复且透明的计算生物医学研究。它允许没有编程经验的用户轻松指定参数并运行单个工具以及更复杂的工作流。同时，它还会记录运行信息，以便任何用户都能重复并理解完整的计算分析过程。\n\n[PathVisio](https:\u002F\u002Fpathvisio.github.io\u002F) 是一款免费的开源通路分析与绘图软件，可用于绘制、编辑和分析生物通路。它使用 Java 开发，并可通过插件进行扩展。\n\n[Orange](https:\u002F\u002Forangedatamining.com\u002F) 是一个功能强大的数据挖掘和机器学习工具包，能够进行数据分析和可视化。\n\n[基本局部比对搜索工具](https:\u002F\u002Fblast.ncbi.nlm.nih.gov\u002FBlast.cgi) 是一种用于寻找生物序列之间相似区域的工具。该程序会将核酸或蛋白质序列与序列数据库进行比较，并计算其统计显著性。\n\n[OSIRIS](https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fosiris\u002F) 是一款公有领域的免费开源 STR 分析软件，专为临床、法医和科研用途而设计，并已被验证可用作针对单来源样本的专家系统。\n\n[NCBI BioSystems](https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fbiosystems\u002F) 是一个数据库，提供对生物系统的集成访问，包括其组成基因、蛋白质和小分子，以及描述这些生物系统的文献和其他相关数据，所有内容均整合在 Entrez 系统中。\n\n# CUDA 开发\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_19a3d218f324.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_34ac80e10409.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n**CUDA 工具包。来源：[NVIDIA 开发者 CUDA](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcuda-zone)**\n\n## CUDA 学习资源\n\n[CUDA](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcuda-zone) 是由 NVIDIA 开发的并行计算平台和编程模型，用于在图形处理器（GPU）上进行通用计算。借助 CUDA，开发者可以充分利用 GPU 的强大算力，显著加速计算应用。在 GPU 加速的应用中，工作负载中的串行部分通常由针对单线程优化的 CPU 来执行，而计算密集型部分则可以在 GPU 的数千个核心上并行运行。使用 CUDA 时，开发者可以采用 C、C++、Fortran、Python 和 MATLAB 等主流编程语言进行开发。\n\n[CUDA 工具包文档](https:\u002F\u002Fdocs.nvidia.com\u002Fcuda\u002Findex.html)\n\n[CUDA 快速入门指南](https:\u002F\u002Fdocs.nvidia.com\u002Fcuda\u002Fcuda-quick-start-guide\u002Findex.html)\n\n[WSL 上的 CUDA](https:\u002F\u002Fdocs.nvidia.com\u002Fcuda\u002Fwsl-user-guide\u002Findex.html)\n\n[TensorFlow 的 CUDA GPU 支持](https:\u002F\u002Fwww.tensorflow.org\u002Finstall\u002Fgpu)\n\n[NVIDIA 深度学习 cuDNN 文档](https:\u002F\u002Fdocs.nvidia.com\u002Fdeeplearning\u002Fcudnn\u002Fapi\u002Findex.html)\n\n[NVIDIA GPU 云文档](https:\u002F\u002Fdocs.nvidia.com\u002Fngc\u002Fngc-introduction\u002Findex.html)\n\n[NVIDIA NGC](https:\u002F\u002Fngc.nvidia.com\u002F) 是一个面向深度学习、机器学习和高性能计算（HPC）工作loads的 GPU 优化软件中心。\n\n[NVIDIA NGC 容器](https:\u002F\u002Fwww.nvidia.com\u002Fen-us\u002Fgpu-cloud\u002Fcontainers\u002F) 是一个注册表，为研究人员、数据科学家和开发者提供对 AI、机器学习和 HPC 领域 GPU 加速软件的全面目录的便捷访问。这些容器能够充分利用本地和云端的 NVIDIA GPU。\n\n## CUDA 工具、库和框架\n\n[CUDA 工具包](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcuda-downloads) 是一套工具和库的集合，为开发高性能 GPU 加速应用提供了完整的开发环境。通过 CUDA 工具包，您可以在 GPU 加速的嵌入式系统、桌面工作站、企业数据中心、云平台以及 HPC 超级计算机上开发、优化和部署您的应用程序。该工具包包含 GPU 加速的库、调试和优化工具、C\u002FC++ 编译器以及运行时库，支持在包括 x86、Arm 和 POWER 在内的主流架构上构建和部署应用。\n\n[NVIDIA cuDNN](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcudnn) 是一个用于 [深度神经网络](https:\u002F\u002Fdeveloper.nvidia.com\u002Fdeep-learning) 的 GPU 加速原语库。cuDNN 提供了针对标准操作的高度优化实现，例如前向和反向卷积、池化、归一化以及激活层。cuDNN 可以加速广泛使用的深度学习框架，包括 [Caffe2](https:\u002F\u002Fcaffe2.ai\u002F)、[Chainer](https:\u002F\u002Fchainer.org\u002F)、[Keras](https:\u002F\u002Fkeras.io\u002F)、[MATLAB](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fdeep-learning.html)、[MxNet](https:\u002F\u002Fmxnet.incubator.apache.org\u002F)、[PyTorch](https:\u002F\u002Fpytorch.org\u002F) 和 [TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002F)。\n\n[CUDA-X HPC](https:\u002F\u002Fwww.nvidia.com\u002Fen-us\u002Ftechnologies\u002Fcuda-x\u002F) 是一组库、工具、编译器和 API 的集合，旨在帮助开发者解决世界上最具挑战性的问题。CUDA-X HPC 包含专为高性能计算（HPC）设计的高度优化的内核。\n\n[NVIDIA 容器工具包](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fnvidia-docker) 是一套工具和库的集合，使用户能够构建并运行 GPU 加速的 Docker 容器。该工具包包括一个容器运行时 [库](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Flibnvidia-container) 以及用于自动配置容器以利用 NVIDIA GPU 的实用程序。\n\n[Minkowski Engine](https:\u002F\u002Fnvidia.github.io\u002FMinkowskiEngine) 是一个用于稀疏张量的自动微分库。它支持所有标准的神经网络层，如卷积、池化、反池化以及稀疏张量上的广播操作。\n\n[CUTLASS](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fcutlass) 是一套 CUDA C++ 模板抽象，用于在 CUDA 中实现各个层次和规模的高性能矩阵乘法（GEMM）。它采用了与 cuBLAS 类似的分层分解和数据移动策略。\n\n[CUB](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fcub) 是为 CUDA C++ 内核开发者提供的协作式原语库。\n\n[Tensorman](https:\u002F\u002Fgithub.com\u002Fpop-os\u002Ftensorman) 是由 [System76]( https:\u002F\u002Fsystem76.com) 开发的用于轻松管理 TensorFlow 容器的工具。Tensorman 允许 TensorFlow 在一个与系统其他部分隔离的环境中运行。这个虚拟环境可以独立于基础系统运行，从而让您在任何支持 Docker 运行时的 Linux 发行版上使用任意版本的 TensorFlow。\n\n[Numba](https:\u002F\u002Fgithub.com\u002Fnumba\u002Fnumba) 是由 Anaconda, Inc. 赞助的面向 Python 的开源、兼容 NumPy 的优化编译器。它利用 LLVM 编译器项目将 Python 语法转换为机器代码。Numba 可以编译大量以数值计算为中心的 Python 代码，包括许多 NumPy 函数。此外，Numba 还支持循环的自动并行化、生成 GPU 加速代码以及创建 ufunc 和 C 回调函数。\n\n[Chainer](https:\u002F\u002Fchainer.org\u002F) 是一个基于 Python 的深度学习框架，旨在提供灵活性。它提供了基于“定义即执行”方法（动态计算图）的自动微分 API，以及用于构建和训练神经网络的面向对象的高级 API。它还通过 [CuPy](https:\u002F\u002Fgithub.com\u002Fcupy\u002Fcupy) 支持 CUDA\u002FcuDNN，以实现高性能的训练和推理。\n\n[CuPy](https:\u002F\u002Fcupy.dev\u002F) 是一个基于 CUDA 的兼容 NumPy 多维数组实现。CuPy 由核心多维数组类 cupy.ndarray 及其上的许多函数组成。它支持 numpy.ndarray 接口的一部分功能。\n\n[CatBoost](https:\u002F\u002Fcatboost.ai\u002F) 是一个快速、可扩展且高性能的 [梯度提升](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FGradient_boosting) 树库，适用于排名、分类、回归及其他机器学习任务，支持 Python、R、Java 和 C++。它同时支持 CPU 和 GPU 上的计算。\n\n[cuDF](https:\u002F\u002Frapids.ai\u002F) 是一个 GPU DataFrame 库，用于加载、连接、聚合、过滤以及对数据进行其他操作。cuDF 提供了一个类似 pandas 的 API，数据工程师和数据科学家会感到非常熟悉，因此他们可以使用它来轻松加速工作流程，而无需深入了解 CUDA 编程。\n\n[cuML](https:\u002F\u002Fgithub.com\u002Frapidsai\u002Fcuml) 是一系列库的集合，实现了机器学习算法和数学原语函数，这些库与其他 RAPIDS 项目具有兼容的 API。cuML 使数据科学家、研究人员和软件工程师能够在不深入 CUDA 编程的情况下，在 GPU 上运行传统的表格型机器学习任务。在大多数情况下，cuML 的 Python API 与 scikit-learn 的 API 相匹配。\n\n[ArrayFire](https:\u002F\u002Farrayfire.com\u002F) 是一个通用库，简化了针对并行和大规模并行架构（包括 CPU、GPU 和其他硬件加速设备）开发软件的过程。\n\n[Thrust](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fthrust) 是一个 C++ 并行编程库，其接口类似于 C++ 标准库。Thrust 的高级别接口极大地提高了程序员的生产力，同时实现了 GPU 和多核 CPU 之间的性能可移植性。\n\n[AresDB](https:\u002F\u002Feng.uber.com\u002Faresdb\u002F) 是一个基于 GPU 的实时分析存储和查询引擎。它具有低查询延迟、高数据新鲜度以及高效的内存和磁盘存储管理能力。\n\n[Arraymancer](https:\u002F\u002Fmratsim.github.io\u002FArraymancer\u002F) 是 Nim 语言中的一个张量（N 维数组）项目。其主要目标是提供一个快速且符合人体工学的 CPU、CUDA 和 OpenCL ndarray 库，以此为基础构建科学计算生态系统。\n\n[Kintinuous](https:\u002F\u002Fgithub.com\u002Fmp3guy\u002FKintinuous) 是一个实时稠密视觉 SLAM 系统，仅需一个低成本的 RGB-D 传感器，即可在数百米范围内实时生成高质量、全局一致的点云和网格重建结果。\n\n[GraphVite](https:\u002F\u002Fgraphvite.io\u002F) 是一个通用图嵌入引擎，专注于在各种应用中实现高速、大规模的嵌入学习。\n\n# MATLAB 开发\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_5bde272346f3.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## MATLAB 学习资源\n\n[MATLAB](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmatlab.html) 是一种编程语言，能够直接进行矩阵和数组数学运算等数值计算。\n\n[MATLAB 文档](https:\u002F\u002Fwww.mathworks.com\u002Fhelp\u002Fmatlab\u002F)\n\n[MATLAB 入门](https:\u002F\u002Fwww.mathworks.com\u002Fhelp\u002Fmatlab\u002Fgetting-started-with-matlab.html)\n\n[MATLAB Academy 提供的 MATLAB 和 Simulink 培训](https:\u002F\u002Fmatlabacademy.mathworks.com)\n\n[MathWorks 认证项目](https:\u002F\u002Fwww.mathworks.com\u002Fservices\u002Ftraining\u002Fcertification.html)\n\n[Udemy 上的 MATLAB 在线课程](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fmatlab\u002F)\n\n[Coursera 上的 MATLAB 在线课程](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=matlab)\n\n[edX 上的 MATLAB 在线课程](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fmatlab)\n\n[构建 MATLAB GUI](https:\u002F\u002Fwww.mathworks.com\u002Fdiscovery\u002Fmatlab-gui.html)\n\n[MATLAB 风格指南 2.0](https:\u002F\u002Fwww.mathworks.com\u002Fmatlabcentral\u002Ffileexchange\u002F46056-matlab-style-guidelines-2-0)\n\n[使用 MATLAB 和 Simulink 设置 Git 版本控制](https:\u002F\u002Fwww.mathworks.com\u002Fhelp\u002Fmatlab\u002Fmatlab_prog\u002Fset-up-git-source-control.html)\n\n[使用 MATLAB 和 Simulink 通过 Git 拉取、推送和获取文件](https:\u002F\u002Fwww.mathworks.com\u002Fhelp\u002Fmatlab\u002Fmatlab_prog\u002Fpush-and-fetch-with-git.html)\n\n[使用 MATLAB 和 Simulink 创建新仓库](https:\u002F\u002Fwww.mathworks.com\u002Fhelp\u002Fmatlab\u002Fmatlab_prog\u002Fadd-folder-to-source-control.html)\n\n[PRMLT](http:\u002F\u002Fprml.github.io\u002F) 是用于 PRML 教材中机器学习算法的 MATLAB 代码。\n\n## MATLAB 工具、库和框架\n\n**[MATLAB 和 Simulink 服务与应用列表](https:\u002F\u002Fwww.mathworks.com\u002Fproducts.html)**\n\n[MATLAB 云](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fcloud.html) 是一项服务，允许您在从 [MathWorks Cloud](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fcloud.html#browser) 到包括 [AWS](https:\u002F\u002Faws.amazon.com\u002F) 和 [Azure](https:\u002F\u002Fazure.microsoft.com\u002F) 在内的 [公有云](https:\u002F\u002Fwww.mathworks.com\u002Fsolutions\u002Fcloud.html#public-cloud) 的各种云环境中运行。\n\n[MATLAB Online™](https:\u002F\u002Fmatlab.mathworks.com) 是一项服务，允许用户通过 Google Chrome 等网页浏览器使用 MATLAB 和 Simulink。\n\n[Simulink](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fsimulink.html) 是一个基于模型的设计的框图环境。它支持嵌入式系统的仿真、自动代码生成和持续测试。\n\n[Simulink Online™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fsimulink-online.html) 是一项通过您的网页浏览器提供 Simulink 访问权限的服务。\n\n[MATLAB Drive™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmatlab-drive.html) 是一项服务，使您能够随时随地存储、访问和处理文件。\n\n[MATLAB 并行服务器™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmatlab-parallel-server.html) 是一种工具，可将 MATLAB® 程序和 Simulink® 仿真扩展到集群和云环境。您可以在桌面端对程序和仿真进行原型设计，然后无需重新编码即可在集群和云上运行。MATLAB 并行服务器支持批处理作业、交互式并行计算以及大型矩阵的分布式计算。\n\n[MATLAB Schemer](https:\u002F\u002Fgithub.com\u002Fscottclowe\u002Fmatlab-schemer) 是一个 MATLAB 软件包，可以轻松更改 MATLAB 显示界面和 GUI 的配色方案（主题）。\n\n[LRSLibrary](https:\u002F\u002Fgithub.com\u002Fandrewssobral\u002Flrslibrary) 是一个用于视频背景建模和减除的低秩与稀疏工具库。该库最初是为视频中的运动目标检测而设计的，但也可用于其他计算机视觉和机器学习问题。\n\n[图像处理工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fimage.html) 是一种工具，提供了一套全面的参考标准算法和工作流应用程序，用于图像处理、分析、可视化以及算法开发。您可以执行图像分割、图像增强、降噪、几何变换、图像配准以及三维图像处理。\n\n[计算机视觉工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fcomputer-vision.html) 是一种工具，提供用于设计和测试计算机视觉、3D 视觉和视频处理系统的算法、函数和应用程序。您可以进行目标检测与跟踪，以及特征检测、提取和匹配。您还可以自动化单目、双目和鱼眼相机的标定流程。对于 3D 视觉，该工具箱支持视觉和点云 SLAM、立体视觉、运动恢复结构以及点云处理。\n\n[统计与机器学习工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fstatistics.html) 是一种工具，提供用于描述、分析和建模数据的函数和应用程序。您可以使用描述性统计、可视化和聚类进行探索性数据分析；拟合概率分布到数据；生成随机数用于蒙特卡洛模拟，并进行假设检验。回归和分类算法使您能够从数据中得出推论，并以交互方式使用分类和回归学习器应用程序，或以编程方式使用 AutoML 构建预测模型。\n\n[Lidar 工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Flidar.html) 是一种工具，提供用于设计、分析和测试激光雷达处理系统的算法、函数和应用程序。您可以执行目标检测与跟踪、语义分割、形状拟合、激光雷达配准以及障碍物检测。Lidar 工具箱支持结合计算机视觉和激光雷达处理的工作流程中的激光雷达与相机交叉标定。\n\n[地图工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmapping.html) 是一种工具，提供用于转换地理数据和创建地图显示的算法和函数。您可以将数据置于地理背景下进行可视化，利用超过 60 种地图投影创建地图显示，并将来自不同来源的数据转换为一致的地理坐标系统。\n\n[UAV 工具箱](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fuav.html) 是一种应用程序，提供用于设计、仿真、测试和部署无人机及无人驾驶飞行器应用的工具和参考应用。您可以设计自主飞行算法、无人机任务和飞行控制器。飞行日志分析器应用程序允许您以交互方式分析常见飞行日志格式中的三维飞行路径、遥测信息和传感器读数。\n\n[并行计算工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmatlab-parallel-server.html) 是一种工具，使您能够利用多核处理器、GPU 和计算机集群来解决计算密集型和数据密集型问题。并行 for 循环、特殊数组类型以及并行化的数值算法等高级构造，使您无需进行 CUDA 或 MPI 编程即可实现 MATLAB® 应用程序的并行化。该工具箱允许您在 MATLAB 及其他工具箱中使用支持并行计算的函数。您还可以将该工具箱与 Simulink® 配合使用，以并行运行模型的多个仿真。程序和模型既可以在交互模式下运行，也可以在批处理模式下运行。\n\n[偏微分方程工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fpde.html) 是一种提供用于通过有限元分析求解结构力学、传热以及一般偏微分方程 (PDE) 的工具。\n\n[ROS 工具箱](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fros.html) 是一种提供 MATLAB® 和 Simulink® 与机器人操作系统 (ROS 和 ROS 2) 之间接口的工具，使您能够创建 ROS 节点网络。该工具箱包含 MATLAB 函数和 Simulink 模块，用于导入、分析和回放以 rosbag 文件格式记录的 ROS 数据。您还可以连接到实时 ROS 网络以访问 ROS 消息。\n\n[机器人技术工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Frobotics.html) 提供一个将机器人技术特定功能（设计、仿真和测试机械臂、移动机器人和人形机器人）引入 MATLAB 的工具箱，充分利用 MATLAB 的原生能力（线性代数、可移植性、图形）。该工具箱还为移动机器人提供了机器人运动模型（自行车模型）、路径规划算法（bug 算法、距离变换、D*、PRM）、运动学动力学规划（格子规划、RRT）、定位（EKF、粒子滤波器）、地图构建（EKF）以及同时定位与地图构建（EKF）等功能，并包含一个非完整约束车辆的 Simulink 模型。此外，该工具箱还包括一个详细的四旋翼飞行机器人 Simulink 模型。\n\n[深度学习工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdeep-learning.html) 是一种提供框架的工具，用于借助算法、预训练模型和应用程序来设计和实现深度神经网络。您可以使用卷积神经网络 (ConvNets, CNNs) 和长短期记忆 (LSTM) 网络对图像、时间序列和文本数据进行分类和回归。借助自动微分、自定义训练循环和共享权重，您可以构建生成对抗网络 (GANs) 和暹罗网络等网络架构。通过 Deep Network Designer 应用程序，您可以以图形化方式设计、分析和训练网络。它可以通过 ONNX 格式与 TensorFlow™ 和 PyTorch 进行模型交换，并可从 TensorFlow-Keras 和 Caffe 导入模型。该工具箱支持迁移学习，提供 DarkNet-53、ResNet-50、NASNet、SqueezeNet 等多种预训练模型。\n\n[强化学习工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Freinforcement-learning.html) 是一种提供应用程序、函数和 Simulink® 块的工具，用于使用强化学习算法（包括 DQN、PPO、SAC 和 DDPG）训练策略。您可以使用这些策略为资源分配、机器人技术和自主系统等复杂应用实现控制器和决策算法。\n\n[深度学习 HDL 工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fdeep-learning-hdl.html) 是一种提供函数和工具的工具，用于在 FPGA 和 SoC 上原型化和实现深度学习网络。它为在受支持的 Xilinx® 和 Intel® FPGA 以及 SoC 设备上运行各种深度学习网络提供了预构建的比特流。性能分析和估算工具使您能够通过探索设计、性能和资源利用率之间的权衡来定制深度学习网络。\n\n[模型预测控制工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fmodel-predictive-control.html) 是一种提供函数、应用程序和 Simulink® 块的工具，用于使用线性和非线性模型预测控制 (MPC) 设计和仿真控制器。该工具箱允许您指定被控对象和干扰模型、预测时域、约束条件和权重。通过运行闭环仿真，您可以评估控制器的性能。\n\n[Vision HDL 工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fvision-hdl.html) 是一种提供像素流算法的工具，用于在 FPGA 和 ASIC 上设计和实现视觉系统。它提供一个设计框架，支持多种接口类型、帧尺寸和帧率。该工具箱中的图像处理、视频和计算机视觉算法采用适合 HDL 实现的架构。\n\n[SoC 块集™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fsoc.html) 是一种提供 Simulink® 块和可视化工具的工具，用于建模、仿真和分析 ASIC、FPGA 以及片上系统 (SoC) 的硬件和软件架构。您可以通过内存模型、总线模型和 I\u002FO 模型构建系统架构，并将架构与算法一起进行仿真。\n\n[无线 HDL 工具箱™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fwireless-hdl.html) 是一种提供经过预先验证、可直接用于硬件实现的 Simulink® 块和子系统的工具，用于开发 5G、LTE 以及基于 OFDM 的自定义无线通信应用。它包括参考应用、IP 模块以及帧级与样本级处理之间的网关。\n\n[ThingSpeak™](https:\u002F\u002Fwww.mathworks.com\u002Fproducts\u002Fthingspeak.html) 是一种物联网数据分析服务，使您能够在云端聚合、可视化和分析实时数据流。ThingSpeak 可为您提供由您的设备发布到 ThingSpeak 的数据的即时可视化效果。借助 ThingSpeak 中执行 MATLAB® 代码的功能，您可以对传入的数据进行在线分析和处理。ThingSpeak 经常用于需要数据分析的物联网系统原型设计和概念验证。\n\n[SEA-MAT](https:\u002F\u002Fsea-mat.github.io\u002Fsea-mat\u002F) 是一项协作项目，旨在为海洋学界组织和分发 MATLAB 工具。\n\n[Gramm](https:\u002F\u002Fgithub.com\u002Fpiermorel\u002Fgramm) 是一个完整的 MATLAB 数据可视化工具箱。它提供易于使用且高层次的界面，可用于生成具有丰富统计可视化效果的高出版质量图表，以展示复杂数据。Gramm 的灵感来源于 R 语言的 ggplot2 库。\n\n[hctsa](https:\u002F\u002Fhctsa-users.gitbook.io\u002Fhctsa-manual) 是一个使用 MATLAB 运行高度比较性时间序列分析的软件包。\n\n[Plotly](https:\u002F\u002Fplot.ly\u002Fmatlab\u002F) 是 MATLAB 的绘图库。\n\n[YALMIP](https:\u002F\u002Fyalmip.github.io\u002F) 是一个用于优化建模的 MATLAB 工具箱。\n\n[GNU Octave](https:\u002F\u002Fwww.gnu.org\u002Fsoftware\u002Foctave\u002F) 是一种高级解释型语言，主要用于数值计算。它提供了求解线性和非线性问题的数值方法，并可用于进行其他数值实验。此外，Octave 还具备强大的绘图功能，可用于数据可视化和处理。\n\n\n\n# C\u002FC++ 开发\n\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_24148fc60257.png\">\n  \u003Cbr \u002F>\n  \n\u003C\u002Fp>\n\n## C\u002FC++ 学习资源\n\n[C++](https:\u002F\u002Fwww.cplusplus.com\u002Fdoc\u002Ftutorial\u002F) 是一种跨平台语言，由 Bjarne Stroustrup 在 C 语言的基础上扩展而来，可用于构建高性能应用程序。\n\n[C](https:\u002F\u002Fwww.iso.org\u002Fstandard\u002F74528.html) 是一种通用的高级编程语言，最初由 Dennis M. Ritchie 在贝尔实验室开发，用于编写 UNIX 操作系统。它支持结构化编程、词法作用域和递归，并采用静态类型系统。C 语言还提供了能够高效映射到典型机器指令的构造，因此成为当今使用最广泛的编程语言之一。\n\n[嵌入式 C](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FEmbedded_C) 是由 [C 标准委员会](https:\u002F\u002Fisocpp.org\u002Fstd\u002Fthe-committee) 为解决不同 [嵌入式系统](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FEmbedded_system) 之间存在的 C 语言扩展差异而制定的一组语言扩展。这些扩展有助于增强微处理器的特性，例如定点运算、多个独立的存储器区域以及基本的 I\u002FO 操作。这使得嵌入式 C 成为全球最受欢迎的嵌入式软件编程语言。\n\n[JetBrains 的 C 和 C++ 开发工具](https:\u002F\u002Fwww.jetbrains.com\u002Fcpp\u002F)\n\n[cppreference.com 上的开源 C++ 库](https:\u002F\u002Fen.cppreference.com\u002Fw\u002Fcpp\u002Flinks\u002Flibs)\n\n[C++ 图形库](https:\u002F\u002Fcpp.libhunt.com\u002Flibs\u002Fgraphics)\n\n[MATLAB 中的 C++ 库](https:\u002F\u002Fwww.mathworks.com\u002Fhelp\u002Fmatlab\u002Fcall-cpp-library-functions.html)\n\n[C++ 工具与库相关文章](https:\u002F\u002Fwww.cplusplus.com\u002Farticles\u002Ftools\u002F)\n\n[Google C++ 风格指南](https:\u002F\u002Fgoogle.github.io\u002Fstyleguide\u002Fcppguide.html)\n\n[Google Developers 上的 C++ 入门课程](https:\u002F\u002Fdevelopers.google.com\u002Fedu\u002Fc++\u002F)\n\n[Fuchsia 的 C++ 风格指南](https:\u002F\u002Ffuchsia.dev\u002Ffuchsia-src\u002Fdevelopment\u002Flanguages\u002Fc-cpp\u002Fcpp-style)\n\n[OpenTitan 的 C 和 C++ 编码规范](https:\u002F\u002Fdocs.opentitan.org\u002Fdoc\u002Frm\u002Fc_cpp_coding_style\u002F)\n\n[Chromium 的 C++ 风格指南](https:\u002F\u002Fchromium.googlesource.com\u002Fchromium\u002Fsrc\u002F+\u002Fmaster\u002Fstyleguide\u002Fc++\u002Fc++.md)\n\n[C++ 核心指南](https:\u002F\u002Fgithub.com\u002Fisocpp\u002FCppCoreGuidelines\u002Fblob\u002Fmaster\u002FCppCoreGuidelines.md)\n\n[ROS 的 C++ 风格指南](http:\u002F\u002Fwiki.ros.org\u002FCppStyleGuide)\n\n[学习 C++](https:\u002F\u002Fwww.learncpp.com\u002F)\n\n[学习 C：交互式 C 教程](https:\u002F\u002Fwww.learn-c.org\u002F)\n\n[C++ 协会](https:\u002F\u002Fcppinstitute.org\u002Ffree-c-and-c-courses)\n\n[LinkedIn Learning 上的 C++ 在线课程](https:\u002F\u002Fwww.linkedin.com\u002Flearning\u002Ftopics\u002Fc-plus-plus)\n\n[W3Schools 上的 C++ 教程](https:\u002F\u002Fwww.w3schools.com\u002Fcpp\u002Fdefault.asp)\n\n[edX 上的 C 语言在线课程](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fc-programming)\n\n[edX 上的 C++ 在线课程](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fc-plus-plus)\n\n[Codecademy 上的 C++ 学习课程](https:\u002F\u002Fwww.codecademy.com\u002Flearn\u002Flearn-c-plus-plus)\n\n[Coursera 上的“人人皆可编程”：C 和 C++ 课程](https:\u002F\u002Fwww.coursera.org\u002Fspecializations\u002Fcoding-for-everyone)\n\n[Coursera 上的“面向程序员的 C++”课程](https:\u002F\u002Fwww.coursera.org\u002Flearn\u002Fc-plus-plus-a)\n\n[Coursera 上的热门 C 课程](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=c%20programming)\n\n[Udemy 上的 C++ 在线课程](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fc-plus-plus\u002F)\n\n[Udemy 上的热门 C 课程](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fc-programming\u002F)\n\n[Udemy 上的“面向初学者的嵌入式 C 编程基础”课程](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Fembedded-c-programming-for-embedded-systems\u002F)\n\n[Udacity 上的“面向程序员的 C++”课程](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fc-for-programmers--ud210)\n\n[Pluralsight 上的“学习 C++ 程序设计”课程](https:\u002F\u002Fwww.pluralsight.com\u002Fcourses\u002Flearn-program-cplusplus)\n\n[MIT 免费在线课程材料中的 C++ 入门](https:\u002F\u002Focw.mit.edu\u002Fcourses\u002Felectrical-engineering-and-computer-science\u002F6-096-introduction-to-c-january-iap-2011\u002F)\n\n[哈佛大学的“面向程序员的 C++ 入门”课程](https:\u002F\u002Fonline-learning.harvard.edu\u002Fcourse\u002Fintroduction-c-programmers)\n\n[哈佛大学的在线 C 课程](https:\u002F\u002Fonline-learning.harvard.edu\u002Fsubject\u002Fc)\n\n\n## C\u002FC++ 工具\n\n[AWS SDK for C++](https:\u002F\u002Faws.amazon.com\u002Fsdk-for-cpp\u002F)\n\n[Azure SDK for C++](https:\u002F\u002Fgithub.com\u002FAzure\u002Fazure-sdk-for-cpp)\n\n[Azure SDK for C](https:\u002F\u002Fgithub.com\u002FAzure\u002Fazure-sdk-for-c)\n\n[Google Cloud 服务的 C++ 客户端库](https:\u002F\u002Fgithub.com\u002Fgoogleapis\u002Fgoogle-cloud-cpp)\n\n[Visual Studio](https:\u002F\u002Fvisualstudio.microsoft.com\u002F) 是微软推出的一款集成开发环境（IDE），功能强大，适用于软件开发的各个方面。Visual Studio 可以轻松地编辑、调试、构建和发布应用程序，尤其适合使用 Windows API、Windows Forms、Windows Presentation Foundation 和 Windows Store 等微软开发平台的项目。\n\n[Visual Studio Code](https:\u002F\u002Fcode.visualstudio.com\u002F) 是一款专为构建和调试现代 Web 和云应用而重新设计和优化的代码编辑器。\n\n[Vcpkg](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fvcpkg) 是一个适用于 Windows、Linux 和 macOS 的 C++ 库管理工具。\n\n[ReSharper C++](https:\u002F\u002Fwww.jetbrains.com\u002Fresharper-cpp\u002Ffeatures\u002F) 是 JetBrains 为 C++ 开发者提供的 Visual Studio 扩展。\n\n[AppCode](https:\u002F\u002Fwww.jetbrains.com\u002Fobjc\u002F) 会持续监控代码质量，及时提醒错误和潜在问题，并自动提供快速修复建议。AppCode 提供针对 Objective-C、Swift、C\u002FC++ 以及其他支持语言的大量代码检查功能，所有检查均实时运行。\n\n[CLion](https:\u002F\u002Fwww.jetbrains.com\u002Fclion\u002Ffeatures\u002F) 是 JetBrains 推出的一款跨平台 C 和 C++ 集成开发环境。\n\n[Code::Blocks](https:\u002F\u002Fwww.codeblocks.org\u002F) 是一款免费的 C\u002FC++ 和 Fortran 集成开发环境，旨在满足用户最苛刻的需求。它具有高度的可扩展性和完全可配置性，基于插件框架设计，可通过安装插件进行功能扩展。\n\n[CppSharp](https:\u002F\u002Fgithub.com\u002Fmono\u002FCppSharp) 是一套工具和库，用于简化原生 C\u002FC++ 代码与 .NET 生态系统的集成。它能够读取 C\u002FC++ 头文件和库文件，并生成必要的胶水代码，将原生 API 封装为托管 API。通过这种方式，可以在托管代码中调用现有的原生库，或为原生代码库添加托管脚本支持。\n\n[Conan](https:\u002F\u002Fconan.io\u002F) 是一个开源的包管理器，专为 C++ 开发和依赖管理而设计，使其步入 21 世纪，并与其他开发生态系统保持同步。\n\n[高性能计算（HPC）SDK](https:\u002F\u002Fdeveloper.nvidia.com\u002Fhpc) 是一套全面的工具箱，用于加速 GPU 上的 HPC 建模和仿真应用。它包含在 NVIDIA 平台上开发 HPC 应用所需的 C、C++ 和 Fortran 编译器、库以及分析工具。\n\n[Thrust](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fthrust) 是一个 C++ 并行编程库，其接口与 C++ 标准库类似。Thrust 的高级别接口极大地提高了程序员的生产力，同时实现了 GPU 和多核 CPU 之间的性能可移植性。它与 CUDA、TBB 和 OpenMP 等成熟技术的互操作性使其能够轻松集成到现有软件中。\n\n[Boost](https:\u002F\u002Fwww.boost.org\u002F) 是一个专注于前沿 C++ 技术的教育平台。自 2007 年以来，Boost 一直参与谷歌年度夏季代码项目，学生们通过参与 Boost 库的开发来提升自己的技能。\n\n[Automake](https:\u002F\u002Fwww.gnu.org\u002Fsoftware\u002Fautomake\u002F) 是一个用于自动生成符合 GNU 编码标准的 Makefile.in 文件的工具。Automake 需要配合使用 GNU Autoconf。\n\n[Cmake](https:\u002F\u002Fcmake.org\u002F) 是一个开源的跨平台工具集，用于构建、测试和打包软件。CMake 使用简单且与平台和编译器无关的配置文件来控制软件的编译过程，并生成可在您选择的编译环境中使用的原生 makefile 和工作区。\n\n[GDB](http:\u002F\u002Fwww.gnu.org\u002Fsoftware\u002Fgdb\u002F) 是一个调试器，它允许您在程序运行时查看其内部状态，或在程序崩溃时了解其当时的执行情况。\n\n[GCC](https:\u002F\u002Fgcc.gnu.org\u002F) 是一个编译器集合，包含 C、C++、Objective-C、Fortran、Ada、Go 和 D 语言的前端，以及这些语言的库。\n\n[GSL](https:\u002F\u002Fwww.gnu.org\u002Fsoftware\u002Fgsl\u002F) 是一个面向 C 和 C++ 程序员的数值库。它是根据 GNU 通用公共许可证发布的自由软件。该库提供了广泛的数学例程，如随机数生成器、特殊函数和最小二乘拟合等。整个库共有超过 1000 个函数，并附带一个庞大的测试套件。\n\n[OpenGL 扩展加载库（GLEW）](https:\u002F\u002Fwww.opengl.org\u002Fsdk\u002Flibs\u002FGLEW\u002F) 是一个跨平台的开源 C\u002FC++ 扩展加载库。GLEW 提供高效的运行时机制，用于确定目标平台上支持哪些 OpenGL 扩展。\n\n[Libtool](https:\u002F\u002Fwww.gnu.org\u002Fsoftware\u002Flibtool\u002F) 是一个通用的库支持脚本，它通过一致且可移植的接口隐藏了使用共享库的复杂性。要使用 Libtool，只需将新的通用库构建命令添加到您的 Makefile、Makefile.in 或 Makefile.am 中即可。\n\n[Maven](https:\u002F\u002Fmaven.apache.org\u002F) 是一个用于软件项目管理和理解的工具。基于项目对象模型（POM）的概念，Maven 可以从一个中心化的信息源管理项目的构建、报告和文档。\n\n[TAU（调优与分析工具）](http:\u002F\u002Fwww.cs.uoregon.edu\u002Fresearch\u002Ftau\u002Fhome.php) 能够通过插桩函数、方法、基本块和语句，以及基于事件的采样来收集性能信息。它支持所有 C++ 语言特性，包括模板和命名空间。\n\n[Clang](https:\u002F\u002Fclang.llvm.org\u002F) 是一个生产级的 C、Objective-C、C++ 和 Objective-C++ 编译器，适用于 X86-32、X86-64 和 ARM 架构（其他目标可能存在一些限制，但通常容易解决）。Clang 已被广泛应用于生产环境，用于构建性能关键型软件，如 Google Chrome 或 Firefox。\n\n[OpenCV](https:\u002F\u002Fopencv.org\u002F) 是一个高度优化的库，专注于实时应用。它提供跨平台的 C++、Python 和 Java 接口，支持 Linux、MacOS、Windows、iOS 和 Android 系统。\n\n[Libcu++](https:\u002F\u002Fnvidia.github.io\u002Flibcudacxx) 是 NVIDIA 为整个系统提供的 C++ 标准库。它提供了 C++ 标准库的异构实现，可以在 CPU 和 GPU 代码中以及两者之间使用。\n\n[ANTLR（另一种语言识别工具）](https:\u002F\u002Fwww.antlr.org\u002F) 是一个功能强大的解析器生成器，可用于读取、处理、执行或转换结构化文本或二进制文件。它被广泛用于构建语言、工具和框架。ANTLR 可以根据文法生成解析器，构建语法树，并生成监听器接口，从而方便对感兴趣短语的识别做出响应。\n\n[Oat++](https:\u002F\u002Foatpp.io\u002F) 是一个轻量级且功能强大的 C++ Web 框架，适用于高度可扩展且资源高效的 Web 应用程序。它没有依赖项，易于移植。\n\n[JavaCPP](https:\u002F\u002Fgithub.com\u002Fbytedeco\u002Fjavacpp) 是一个程序，它能够在 Java 中高效地访问原生 C++ 代码，类似于某些 C\u002FC++ 编译器与汇编语言交互的方式。\n\n[Cython](https:\u002F\u002Fcython.org\u002F) 是一种语言，它使得为 Python 编写 C 扩展变得像编写 Python 代码一样简单。Cython 基于 Pyrex，但支持更多前沿功能和优化，例如调用 C 函数以及在变量和类属性上声明 C 类型。\n\n[Spdlog](https:\u002F\u002Fgithub.com\u002Fgabime\u002Fspdlog) 是一个非常快速、仅包含头文件或已编译的 C++ 日志记录库。\n\n[Infer](https:\u002F\u002Ffbinfer.com\u002F) 是一个用于 Java、C++、Objective-C 和 C 语言的静态分析工具。Infer 是用 [OCaml](https:\u002F\u002Focaml.org\u002F) 编写的。\n\n\n# Java 开发\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_d898b973591b.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n## Java 学习资源\n\n[Java](https:\u002F\u002Fwww.oracle.com\u002Fjava\u002F) 是一种流行的编程语言和开发平台（JDK）。它能够降低成本、缩短开发周期、推动创新并提升应用服务。全球有数百万开发者运行着超过 510 亿个 Java 虚拟机。\n\n[Eclipse 基金会](https:\u002F\u002Fwww.eclipse.org\u002Fdownloads\u002F) 汇聚了全球开发者社区，提供 Eclipse IDE、Jakarta EE 以及超过 375 个开源项目，涵盖 Java 及其他语言的运行时、工具和框架。\n\n[Java 入门教程](https:\u002F\u002Fdocs.oracle.com\u002Fjavase\u002Ftutorial\u002F)\n\n[Oracle University 提供的 Oracle Java 认证](https:\u002F\u002Feducation.oracle.com\u002Fjava-certification-benefits)\n\n[Google Developers 培训](https:\u002F\u002Fdevelopers.google.com\u002Ftraining\u002F)\n\n[Google Developers 认证](https:\u002F\u002Fdevelopers.google.com\u002Fcertification\u002F)\n\n[W3Schools 的 Java 教程](https:\u002F\u002Fwww.w3schools.com\u002Fjava\u002F)\n\n[使用 Java 构建你的第一个 Android 应用](codelabs.developers.google.com\u002Fcodelabs\u002Fbuild-your-first-android-app\u002F)\n\n[在 Visual Studio Code 中开始使用 Java](https:\u002F\u002Fcode.visualstudio.com\u002Fdocs\u002Fjava\u002Fjava-tutorial)\n\n[Google Java 编码规范](https:\u002F\u002Fgoogle.github.io\u002Fstyleguide\u002Fjavaguide.html)\n\n[AOSP 贡献者 Java 代码风格指南](https:\u002F\u002Fsource.android.com\u002Fsetup\u002Fcontribute\u002Fcode-style)\n\n[Chromium Java 风格指南](https:\u002F\u002Fchromium.googlesource.com\u002Fchromium\u002Fsrc\u002F+\u002Fmaster\u002Fstyleguide\u002Fjava\u002Fjava.md)\n\n[开始使用 OR-Tools for Java](https:\u002F\u002Fdevelopers.google.com\u002Foptimization\u002Fintroduction\u002Fjava)\n\n[Azure Pipelines 中的 Java 工具安装任务入门](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fdevops\u002Fpipelines\u002Ftasks\u002Ftool\u002Fjava-tool-installer)\n\n[Gradle 用户手册](https:\u002F\u002Fdocs.gradle.org\u002Fcurrent\u002Fuserguide\u002Fuserguide.html)\n\n## 工具\n\n[Java SE](https:\u002F\u002Fwww.oracle.com\u002Fjava\u002Ftechnologies\u002Fjavase\u002Ftools-jsp.html) 包含多种工具，用于辅助程序开发与调试，以及监控和排查生产环境中的应用问题。\n\n[JDK 开发工具](https:\u002F\u002Fdocs.oracle.com\u002Fjavase\u002F7\u002Fdocs\u002Ftechnotes\u002Ftools\u002F) 包括 Java Web Start 工具 (javaws)、Java 排查、性能分析、监控与管理工具 (jcmd, jconsole, jmc, jvisualvm)，以及 Java Web 服务工具 (schemagen, wsgen, wsimport, xjc)。\n\n[Android Studio](https:\u002F\u002Fdeveloper.android.com\u002Fstudio\u002F) 是 Google Android 操作系统的官方集成开发环境，基于 JetBrains 的 IntelliJ IDEA 软件构建，专为 Android 开发设计。支持 Windows、macOS、Linux 和 Chrome OS。\n\n[IntelliJ IDEA](https:\u002F\u002Fwww.jetbrains.com\u002Fidea\u002F) 是一款面向 Java 的 IDE，但它同样能够理解并提供智能编码辅助功能，适用于 Kotlin、SQL、JPQL、HTML、JavaScript 等多种语言，即使这些语言的表达被嵌入到 Java 代码的字符串字面量中。\n\n[NetBeans](https:\u002F\u002Fnetbeans.org\u002Ffeatures\u002Fjava\u002Findex.html) 是一款 IDE，为 Java 开发者提供了创建专业桌面、移动和企业级应用所需的所有工具。包括创建、编辑和重构等功能。该 IDE 还提供向导和模板，帮助用户快速构建 Java EE、Java SE 和 Java ME 应用程序。\n\n[Java 设计模式](https:\u002F\u002Fgithub.com\u002Filuwatar\u002Fjava-design-patterns) 是一组经过最佳形式化实践的设计模式集合，程序员可以利用它们解决应用程序或系统设计过程中常见的问题。\n\n[Elasticsearch](https:\u002F\u002Fwww.elastic.co\u002Fproducts\u002Felasticsearch) 是一个基于 Java 构建的分布式 RESTful 搜索引擎，专为云环境设计。\n\n[RxJava](https:\u002F\u002Fgithub.com\u002FReactiveX\u002FRxJava) 是 [Reactive Extensions](http:\u002F\u002Freactivex.io\u002F) 在 Java 虚拟机上的实现：一个通过可观察序列来组合异步和事件驱动程序的库。它扩展了 [观察者模式](http:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FObserver_pattern)，以支持数据\u002F事件序列，并添加了一系列操作符，允许你以声明式的方式组合序列，同时屏蔽底层线程、同步、线程安全及并发数据结构等细节。\n\n[Guava](https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fguava) 是 Google 提供的一组核心 Java 类库，包含新的集合类型（如 multimap 和 multiset）、不可变集合、图库，以及用于并发、I\u002FO、哈希、缓存、基本类型、字符串等方面的实用工具！它在 Google 内部的大多数 Java 项目中被广泛使用，同时也被许多其他公司采用。\n\n[OkHttp](https:\u002F\u002Fsquare.github.io\u002Fokhttp\u002F) 是 Square 公司开发的一款适用于 Java 和 Kotlin 的 HTTP 客户端。\n\n[Retrofit](https:\u002F\u002Fsquare.github.io\u002Fretrofit\u002F) 是 Square 公司开发的一款适用于 Android 和 Java 的类型安全 HTTP 客户端。\n\n[LeakCanary](https:\u002F\u002Fsquare.github.io\u002Fleakcanary\u002F) 是 Square 公司开发的一款用于 Android 的内存泄漏检测库。\n\n[Apache Spark](https:\u002F\u002Fspark.apache.org\u002F) 是一个用于大规模数据处理的统一分析引擎。它提供了 Scala、Java、Python 和 R 等高级 API，以及一个优化的执行引擎，支持通用计算图进行数据分析。此外，它还提供一系列高级工具，包括用于 SQL 和 DataFrame 的 Spark SQL、用于机器学习的 MLlib、用于图处理的 GraphX，以及用于流处理的 Structured Streaming。\n\n[Apache Flink](https:\u002F\u002Fflink.apache.org\u002F) 是一个开源的流处理框架，具备强大的流处理和批处理能力，拥有优雅且流畅的 Java 和 Scala API。\n\n[Fastjson](https:\u002F\u002Fgithub.com\u002Falibaba\u002Ffastjson\u002Fwiki) 是一个 Java 库，可用于将 Java 对象转换为 JSON 表示形式，也可以将 JSON 字符串转换为相应的 Java 对象。\n\n[libGDX](https:\u002F\u002Flibgdx.com\u002F) 是一个基于 OpenGL (ES) 的跨平台 Java 游戏开发框架，可在 Windows、Linux、Mac OS X、Android、支持 WebGL 的浏览器以及 iOS 上运行。\n\n[Jenkins](https:\u002F\u002Fwww.jenkins.io\u002F) 是领先的开源自动化服务器。它基于 Java 构建，提供了超过 1700 个 [插件](https:\u002F\u002Fplugins.jenkins.io\u002F) 来支持几乎任何自动化的场景，从而使人类能够将时间投入到机器无法完成的任务上。\n\n[DBeaver](https:\u002F\u002Fdbeaver.io\u002F) 是一款免费的多平台数据库工具，适用于开发者、SQL 程序员、数据库管理员和分析师。它支持任何具有 JDBC 驱动的数据库（基本上意味着——任何数据库）。其企业版还支持非 JDBC 数据源（MongoDB、Cassandra、Redis、DynamoDB 等）。\n\n[Redisson](https:\u002F\u002Fredisson.pro\u002F) 是一个具备内存数据网格功能的 Redis Java 客户端。它提供了超过 50 种基于 Redis 的 Java 对象和服务：Set、Multimap、SortedSet、Map、List、Queue、Deque、Semaphore、Lock、AtomicLong、Map Reduce、发布\u002F订阅、Bloom 过滤器、Spring Cache、Tomcat、Scheduler、JCache API、Hibernate、MyBatis、RPC 以及本地缓存。\n\n[GraalVM](https:\u002F\u002Fwww.graalvm.org\u002F) 是一种通用虚拟机，可用于运行用 JavaScript、Python、Ruby、R 语言以及基于 JVM 的语言（如 Java、Scala、Clojure、Kotlin）和基于 LLVM 的语言（如 C 和 C++）编写的应用程序。\n\n[Gradle](https:\u002F\u002Fgradle.org\u002F) 是一款用于多语言软件开发的构建自动化工具。无论是移动应用还是微服务，无论小型初创公司还是大型企业，Gradle 都能帮助团队更快地构建、自动化并交付更优质的软件。你可以使用 Java、C++、Python 或任何你偏好的语言进行开发。\n\n[Apache Groovy](http:\u002F\u002Fwww.groovy-lang.org\u002F) 是一种功能强大、可选类型且动态的语言，同时具备静态类型检查和静态编译的能力，专为 Java 平台设计，旨在通过简洁、熟悉且易于学习的语法提高开发效率。它能够与任何 Java 程序无缝集成，并立即为你的应用程序带来强大的功能，包括脚本支持、领域特定语言的编写、运行时和编译时元编程以及函数式编程等特性。\n\n[JaCoCo](https:\u002F\u002Fwww.jacoco.org\u002Fjacoco\u002F) 是一个面向 Java 的免费代码覆盖率工具库，由 EclEmma 团队基于多年使用和集成现有库的经验而开发。\n\n[Apache JMeter](http:\u002F\u002Fjmeter.apache.org\u002F) 用于测试静态和动态资源以及 Web 动态应用程序的性能。它还可以模拟对服务器、服务器集群、网络或目标对象的高负载，以测试其承受能力或分析在不同负载类型下的整体性能。\n\n[Junit](https:\u002F\u002Fjunit.org\u002F) 是一个用于编写可重复测试的简单框架。它是 xUnit 架构在单元测试框架中的一个实例。\n\n[Mockito](https:\u002F\u002Fsite.mockito.org\u002F) 是最流行的 Java 单元测试模拟框架。\n\n[SpotBugs](https:\u002F\u002Fspotbugs.github.io\u002F) 是一款利用静态分析技术来查找 Java 代码中潜在缺陷的工具。\n\n[SpringBoot](https:\u002F\u002Fspring.io\u002Fprojects\u002Fspring-boot) 是一款出色的工具，可以帮助你以极低的复杂度创建基于 Spring 的生产级应用程序和服务。它对 Spring 平台采取了一种“有观点”的方法，使新老用户都能快速上手所需的功能。\n\n[YourKit](https:\u002F\u002Fwww.yourkit.com\u002F) 是一家技术领先企业，致力于打造最创新、最智能的 Java 和 .NET 应用程序性能分析工具。\n\n\n\n# Python 开发\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_83a9b576f0a0.png\">\n  \u003Cbr \u002F>\n\n\u003C\u002Fp>\n\n## Python 学习资源\n\n[Python](https:\u002F\u002Fwww.python.org) 是一种解释型的高级编程语言。Python 在数据科学和机器学习领域得到了广泛应用。\n\n[Python 开发者指南](https:\u002F\u002Fdevguide.python.org) 是一份全面的资源，适用于所有希望为 Python 社区做出贡献的新老开发者。该指南由维护 Python 的同一社区负责更新和维护。\n\n[Azure Functions Python 开发者指南](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fazure-functions\u002Ffunctions-reference-python) 是关于如何使用 Python 开发 Azure Functions 的入门指南。以下内容假定你已经阅读过 [Azure Functions 开发者指南](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fazure-functions\u002Ffunctions-reference)。\n\n[CheckiO](https:\u002F\u002Fcheckio.org\u002F) 是一个编程学习平台兼游戏化网站，通过解决代码挑战并争夺最优雅、最具创意的解决方案来教授 Python。\n\n[Python Institute](https:\u002F\u002Fpythoninstitute.org)\n\n[PCEP – 认证初级 Python 程序员资格认证](https:\u002F\u002Fpythoninstitute.org\u002Fpcep-certification-entry-level\u002F)\n\n[PCAP – 认证 Python 编程助理资格认证](https:\u002F\u002Fpythoninstitute.org\u002Fpcap-certification-associate\u002F)\n\n[PCPP – 认证 Python 编程专业人员 1 资格认证](https:\u002F\u002Fpythoninstitute.org\u002Fpcpp-certification-professional\u002F)\n\n[PCPP – 认证 Python 编程专业人员 2 资格认证](https:\u002F\u002Fpythoninstitute.org\u002Fpcpp-certification-professional\u002F)\n\n[MTA：使用 Python 进行编程入门认证](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Flearn\u002Fcertifications\u002Fmta-introduction-to-programming-using-python)\n\n[在 Visual Studio Code 中开始使用 Python](https:\u002F\u002Fcode.visualstudio.com\u002Fdocs\u002Fpython\u002Fpython-tutorial)\n\n[Google 的 Python 风格指南](https:\u002F\u002Fgoogle.github.io\u002Fstyleguide\u002Fpyguide.html)\n\n[Google 的 Python 教育课程](https:\u002F\u002Fdevelopers.google.com\u002Fedu\u002Fpython\u002F)\n\n[Real Python](https:\u002F\u002Frealpython.com)\n\n[Forrest Knight 提供的开源 Python 计算机科学学位课程](https:\u002F\u002Fgithub.com\u002FForrestKnight\u002Fopen-source-cs-python)\n\n[数据科学领域的 Python 入门课程](https:\u002F\u002Fwww.datacamp.com\u002Fcourses\u002Fintro-to-python-for-data-science)\n\n[W3schools 的 Python 入门课程](https:\u002F\u002Fwww.w3schools.com\u002Fpython\u002Fpython_intro.asp)\n\n[Codecademy 的 Python 3 课程](https:\u002F\u002Fwww.codecademy.com\u002Flearn\u002Flearn-python-3)\n\n[通过 edX 的在线课程学习 Python](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fpython)\n\n[Coursera 上的 Python 在线课程](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=python)\n\n## Python 框架与工具\n\n[Python 包索引 (PyPI)](https:\u002F\u002Fpypi.org\u002F) 是 Python 编程语言的软件仓库。PyPI 帮助你查找和安装由 Python 社区开发并共享的软件。\n\n[PyCharm](https:\u002F\u002Fwww.jetbrains.com\u002Fpycharm\u002F) 是我用过的最好的 IDE。借助 PyCharm，你可以在一个地方访问命令行、连接数据库、创建虚拟环境以及管理版本控制系统，从而避免在不同窗口之间频繁切换，节省时间。\n\n[Python Tools for Visual Studio (PTVS)](https:\u002F\u002Fmicrosoft.github.io\u002FPTVS\u002F) 是一个免费的开源插件，可将 Visual Studio 转变为 Python IDE。它支持编辑、浏览、IntelliSense、Python\u002FC++ 混合调试、远程 Linux\u002FMacOS 调试、性能分析、IPython 以及使用 Django 等框架进行 Web 开发。\n\n[Pylance](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fpylance-release) 是一个扩展，与 Visual Studio Code 中的 Python 配合使用，提供高效的语言支持。Pylance 的底层由 Microsoft 的静态类型检查工具 Pyright 提供支持。\n\n[Pyright](https:\u002F\u002Fgithub.com\u002FMicrosoft\u002Fpyright) 是一款针对大型 Python 代码库的快速类型检查器。它可以以“监视”模式运行，并在文件修改时执行快速的增量更新。\n\n[Django](https:\u002F\u002Fwww.djangoproject.com\u002F) 是一个高级 Python Web 框架，鼓励快速开发和简洁、实用的设计。\n\n[Flask](https:\u002F\u002Fflask.palletsprojects.com\u002F) 是一个用 Python 编写的微型 Web 框架。它被归类为微型框架，因为它不需要特定的工具或库。\n\n[Web2py](http:\u002F\u002Fweb2py.com\u002F) 是一个用 Python 编写的开源 Web 应用程序框架，允许 Web 开发人员编写动态 Web 内容。单个 Web2py 实例可以运行多个使用不同数据库的网站。\n\n[AWS Chalice](https:\u002F\u002Fgithub.com\u002Faws\u002Fchalice) 是一个用于用 Python 编写无服务器应用程序的框架。它允许你快速创建和部署使用 AWS Lambda 的应用程序。\n\n[Tornado](https:\u002F\u002Fwww.tornadoweb.org\u002F) 是一个 Python Web 框架和异步网络库。Tornado 使用非阻塞式网络 I\u002FO，能够扩展到数以万计的开放连接。\n\n[HTTPie](https:\u002F\u002Fgithub.com\u002Fhttpie\u002Fhttpie) 是一个命令行 HTTP 客户端，旨在尽可能简化与 Web 服务的 CLI 交互。HTTPie 专为测试、调试以及与 API 和 HTTP 服务器的一般交互而设计。\n\n[Scrapy](https:\u002F\u002Fscrapy.org\u002F) 是一个快速、高级的 Web 爬取和数据抓取框架，用于爬取网站并从其页面中提取结构化数据。它可以用于广泛的目的，从数据挖掘到监控和自动化测试。\n\n[Sentry](https:\u002F\u002Fsentry.io\u002F) 是一项帮助你实时监控和修复崩溃的服务。其服务器端使用 Python 编写，但包含完整的 API，可用于从任何语言、任何应用程序发送事件。\n\n[Pipenv](https:\u002F\u002Fgithub.com\u002Fpypa\u002Fpipenv) 是一种工具，旨在将各个包管理领域的最佳实践（如 Bundler、Composer、npm、Cargo、Yarn 等）引入 Python 生态系统。\n\n[Python Fire](https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fpython-fire) 是一个库，可以从任何 Python 对象自动生成命令行界面 (CLI)。\n\n[Bottle](https:\u002F\u002Fgithub.com\u002Fbottlepy\u002Fbottle) 是一个快速、简单且轻量级的 [WSGI](https:\u002F\u002Fwww.wsgi.org\u002F) 微型 Web 框架，专为 Python 设计。它以单个文件模块的形式分发，除了 [Python 标准库](https:\u002F\u002Fdocs.python.org\u002Flibrary\u002F) 外没有任何依赖。\n\n[CherryPy](https:\u002F\u002Fcherrypy.org) 是一个极简主义的面向对象 Python HTTP Web 框架。\n\n[Sanic](https:\u002F\u002Fgithub.com\u002Fhuge-success\u002Fsanic) 是一个基于 Python 3.6+ 的 Web 服务器和 Web 框架，专为高性能而设计。\n\n[Pyramid](https:\u002F\u002Ftrypyramid.com) 是一个小型且快速的开源 Python Web 框架。它使实际的 Web 应用开发和部署更加有趣和高效。\n\n[TurboGears](https:\u002F\u002Fturbogears.org) 是一个混合型 Web 框架，既可以作为全栈框架使用，也可以作为微型框架使用。\n\n[Falcon](https:\u002F\u002Ffalconframework.org\u002F) 是一个可靠、高性能的 Python Web 框架，用于构建大规模应用程序后端和微服务，支持 MongoDB、可插拔应用以及自动生成的管理员界面。\n\n[Neural Network Intelligence (NNI)](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fnni) 是一个开源 AutoML 工具包，用于自动化机器学习生命周期，包括 [特征工程](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fnni\u002Fblob\u002Fmaster\u002Fdocs\u002Fen_US\u002FFeatureEngineering\u002FOverview.md)、[神经架构搜索](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fnni\u002Fblob\u002Fmaster\u002Fdocs\u002Fen_US\u002FNAS\u002FOverview.md)、[模型压缩](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fnni\u002Fblob\u002Fmaster\u002Fdocs\u002Fen_US\u002FCompressor\u002FOverview.md) 和 [超参数调优](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fnni\u002Fblob\u002Fmaster\u002Fdocs\u002Fen_US\u002FTuner\u002FBuiltinTuner.md)。\n\n[Dash](https:\u002F\u002Fplotly.com\u002Fdash) 是一个流行的 Python 框架，用于为 Python、R、Julia 和 Jupyter 构建机器学习和数据科学 Web 应用程序。\n\n[Luigi](https:\u002F\u002Fgithub.com\u002Fspotify\u002Fluigi) 是一个 Python 模块，可以帮助你构建复杂的批处理作业管道。它负责处理依赖关系解析、工作流管理、可视化等功能，并内置了对 Hadoop 的支持。\n\n[Locust](https:\u002F\u002Fgithub.com\u002Flocustio\u002Flocust) 是一个易于使用、可脚本化且可扩展的性能测试工具。\n\n[spaCy](https:\u002F\u002Fgithub.com\u002Fexplosion\u002FspaCy) 是一个用于 Python 和 Cython 的高级自然语言处理库。\n\n[NumPy](https:\u002F\u002Fwww.numpy.org\u002F) 是使用 Python 进行科学计算所必需的基础包。\n\n[Pillow](https:\u002F\u002Fpython-pillow.org\u002F) 是 PIL（Python 图像库）的一个友好分支。\n\n[IPython](https:\u002F\u002Fipython.org\u002F) 是一种多语言交互式计算的命令行外壳，最初为 Python 编程语言开发，提供了增强的内省功能、富媒体支持、额外的 shell 语法、Tab 补全以及丰富的历史记录。\n\n[GraphLab Create](https:\u002F\u002Fturi.com\u002F) 是一个基于 C++ 引擎的 Python 库，可用于快速构建大规模、高性能的机器学习模型。\n\n[Pandas](https:\u002F\u002Fpandas.pydata.org\u002F) 是一个快速、强大且易于使用的开源数据结构、数据分析和操作工具，构建于 Python 编程语言之上。\n\n[PuLP](https:\u002F\u002Fcoin-or.github.io\u002Fpulp\u002F) 是一个用 Python 编写的线性规划建模工具。PuLP 可以生成 LP 文件，并调用高度优化的求解器（如 GLPK、COIN CLP\u002FCBC、CPLEX 和 GUROBI）来解决这些线性问题。\n\n[Matplotlib](https:\u002F\u002Fmatplotlib.org\u002F) 是一个用于在 Python 中创建静态、动画和交互式可视化效果的 2D 绘图库。Matplotlib 可以在各种打印格式和跨平台的交互式环境中生成出版质量的图表。\n\n[Scikit-Learn](https:\u002F\u002Fscikit-learn.org\u002Fstable\u002Findex.html) 是一个简单高效的数据挖掘和数据分析工具。它构建于 NumPy、SciPy 和 Matplotlib 之上。\n\n# Scala 开发\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_334a63768e64.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n\n# Scala 学习资源\n\n[Scala](https:\u002F\u002Fscala-lang.org\u002F) 是一种将面向对象编程与函数式编程结合于一体的简洁、高级语言。Scala 的静态类型系统有助于避免复杂应用程序中的错误，而其 JVM 和 JavaScript 运行时则使你能够构建高性能系统，并轻松访问庞大的库生态系统。\n\n[Scala 风格指南](https:\u002F\u002Fdocs.scala-lang.org\u002Fstyle\u002F)\n\n[Databricks Scala 风格指南](https:\u002F\u002Fgithub.com\u002Fdatabricks\u002Fscala-style-guide)\n\n[在 Azure 上使用 Scala 和 Spark 进行数据科学](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fmachine-learning\u002Fteam-data-science-process\u002Fscala-walkthrough)\n\n[使用 IntelliJ 在 HDInsight 中为 Apache Spark 创建 Scala Maven 应用程序](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fhdinsight\u002Fspark\u002Fapache-spark-create-standalone-application)\n\n[使用 Scala 和 Azure Databricks 介绍 Spark DataFrames](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Fdatabricks\u002Fspark\u002Flatest\u002Fdataframes-datasets\u002Fintroduction-to-dataframes-scala)\n\n[使用 Scala 编写 AWS Glue ETL 脚本](https:\u002F\u002Fdocs.aws.amazon.com\u002Fglue\u002Flatest\u002Fdg\u002Fglue-etl-scala-using.html)\n\n[在 Amazon EMR 集群上使用 Flink Scala Shell](https:\u002F\u002Fdocs.aws.amazon.com\u002Femr\u002Flatest\u002FReleaseGuide\u002Fflink-scala.html)\n\n[Udemy 上的 AWS EMR 和 Spark 2 使用 Scala 课程](https:\u002F\u002Fwww.udemy.com\u002Fcourse\u002Faws-emr-and-spark-2-using-scala\u002F)\n\n[将 Google Cloud Storage 连接器与 Apache Spark 结合使用](https:\u002F\u002Fcloud.google.com\u002Fdataproc\u002Fdocs\u002Ftutorials\u002Fgcs-connector-spark-tutorial)\n\n[在 Google Cloud 的 Cloud Dataproc 上编写并运行 Spark Scala 作业](https:\u002F\u002Fcloud.google.com\u002Fdataproc\u002Fdocs\u002Ftutorials\u002Fspark-scala)\n\n[edX 上的 Scala 课程和认证](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fscala)\n\n[Coursera 上的 Scala 课程](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=scala)\n\n[Udemy 上的顶级 Scala 课程](https:\u002F\u002Fwww.udemy.com\u002Ftopic\u002Fscala\u002F)\n\n# Scala 工具\n\n[Apache Spark](https:\u002F\u002Fspark.apache.org\u002F) 是一个用于大规模数据处理的统一分析引擎。它提供了 Scala、Java、Python 和 R 等高级 API，以及一个优化的引擎，支持用于数据分析的通用计算图。它还支持丰富的高级工具，包括用于 SQL 和 DataFrame 的 Spark SQL、用于机器学习的 MLlib、用于图处理的 GraphX，以及用于流处理的 Structured Streaming。\n\n[SQL Server 和 Azure SQL 的 Apache Spark 连接器](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fsql-spark-connector) 是一种高性能连接器，使你能够在大数据分析中使用事务性数据，并将结果持久化以供即席查询或报告使用。该连接器允许你将任何 SQL 数据库（无论是在本地还是云端）用作 Spark 作业的输入数据源或输出数据接收端。\n\n[Azure Databricks](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fdatabricks\u002F) 是一种快速且协作式的基于 Apache Spark 的大数据分析服务，专为数据科学和数据工程设计。Azure Databricks 可在几分钟内为你设置好 Apache Spark 环境，自动扩展规模，并在一个交互式工作区中协作完成共享项目。Azure Databricks 支持 Python、Scala、R、Java 和 SQL，以及 TensorFlow、PyTorch 和 scikit-learn 等数据科学框架和库。\n\n[Apache PredictionIO](https:\u002F\u002Fpredictionio.apache.org\u002F) 是一个面向开发者、数据科学家和最终用户的开源机器学习框架。它支持事件收集、算法部署、评估，以及通过 REST API 查询预测结果。它基于 Hadoop、HBase（以及其他数据库）、Elasticsearch 和 Spark 等可扩展的开源服务，并实现了所谓的 Lambda 架构。\n\n[Apache Kafka 集群管理器 (CMAK)](https:\u002F\u002Fgithub.com\u002Fyahoo\u002FCMAK) 是一个用于管理 [Apache Kafka](https:\u002F\u002Fkafka.apache.org\u002F) 集群的工具。\n\n[BigDL](https:\u002F\u002Fbigdl-project.github.io\u002F) 是一个用于 Apache Spark 的分布式深度学习库。借助 BigDL，用户可以将他们的深度学习应用程序编写为标准的 Spark 程序，这些程序可以直接在现有的 Spark 或 Hadoop 集群上运行。\n\n[Eclipse Deeplearning4J (DL4J)](https:\u002F\u002Fdeeplearning4j.konduit.ai\u002F) 是一组旨在支持基于 JVM（Scala、Kotlin、Clojure 和 Groovy）的深度学习应用程序所有需求的项目。这意味着从原始数据开始，无论数据来自何处、采用何种格式，都可以对其进行加载和预处理，进而构建和调整各种简单及复杂的深度学习网络。\n\n[Play Framework](https:\u002F\u002Fgithub.com\u002Fplayframework\u002Fplayframework) 是一个结合了生产力和性能的 Web 框架，使使用 Java 和 Scala 构建可扩展的 Web 应用程序变得容易。\n\n[Dotty](https:\u002F\u002Fgithub.com\u002Flampepfl\u002Fdotty) 是一种研究型编译器，未来将成为 Scala 3。\n\n[AWScala](https:\u002F\u002Fgithub.com\u002Fseratch\u002FAWScala) 是一个工具，使 Scala 开发人员能够以 Scala 的方式轻松地与 Amazon Web Services 集成。\n\n[Scala.js](https:\u002F\u002Fwww.scala-js.org\u002F) 是一个将 Scala 转换为 JavaScript 的编译器。\n\n[Polynote](https:\u002F\u002Fpolynote.org\u002F) 是一个实验性的多语言笔记本环境。目前，它支持 Scala 和 Python（无论是否使用 Spark）、SQL 和 Vega。\n\n[Scala Native](http:\u002F\u002Fscala-native.org\u002F) 是一个专门针对 Scala 设计的优化型提前编译器和轻量级托管运行时。\n\n[Gitbucket](https:\u002F\u002Fgitbucket.github.io\u002F) 是一个由 Scala 驱动的 Git 平台，安装简便、扩展性强，并且兼容 GitHub API。\n\n[Finagle](https:\u002F\u002Ftwitter.github.io\u002Ffinagle) 是一个容错、协议无关的 RPC 系统。\n\n[Gatling](https:\u002F\u002Fgatling.io\u002F) 是一个负载测试工具。它官方支持 HTTP、WebSocket、服务器发送事件和 JMS。\n\n[Scalatra](https:\u002F\u002Fscalatra.org\u002F) 是一个小型的 Scala 高性能异步 Web 框架，灵感来源于 [Sinatra](https:\u002F\u002Fwww.sinatrarb.com\u002F)。\n\n\n# R 开发\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_e3cb2c3b58cd.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n# R 学习资源\n\n[R](https:\u002F\u002Fwww.r-project.org\u002F) 是一个用于统计计算和图形绘制的开源软件环境。它可以在多种平台上编译和运行，例如 Windows 和 macOS。\n\n[R 语言入门](https:\u002F\u002Fcran.r-project.org\u002Fdoc\u002Fmanuals\u002Fr-release\u002FR-intro.pdf)\n\n[Google 的 R 风格指南](https:\u002F\u002Fgoogle.github.io\u002Fstyleguide\u002FRguide.html)\n\n[R 开发者 Azure 指南](https:\u002F\u002Fdocs.microsoft.com\u002Fen-us\u002Fazure\u002Farchitecture\u002Fdata-guide\u002Ftechnology-choices\u002Fr-developers-guide)\n\n[在 Google Compute Engine 上大规模运行 R](https:\u002F\u002Fcloud.google.com\u002Fsolutions\u002Frunning-r-at-scale)\n\n[在 AWS 上运行 R](https:\u002F\u002Faws.amazon.com\u002Fblogs\u002Fbig-data\u002Frunning-r-on-aws\u002F)\n\n[RStudio Server Pro for AWS](https:\u002F\u002Faws.amazon.com\u002Fmarketplace\u002Fpp\u002FRStudio-RStudio-Server-Pro-for-AWS\u002FB06W2G9PRY)\n\n[通过 Codecademy 学习 R](https:\u002F\u002Fwww.codecademy.com\u002Flearn\u002Flearn-r)\n\n[通过 edX 的在线课程和教程学习 R 编程](https:\u002F\u002Fwww.edx.org\u002Flearn\u002Fr-programming)\n\n[Coursera 上的 R 语言课程](https:\u002F\u002Fwww.coursera.org\u002Fcourses?query=r%20language)\n\n[通过 Udacity 学习用于数据科学的 R](https:\u002F\u002Fwww.udacity.com\u002Fcourse\u002Fprogramming-for-data-science-nanodegree-with-R--nd118)\n\n# R 工具\n\n[RStudio](https:\u002F\u002Frstudio.com\u002F) 是一个集成了 R 和 Python 的开发环境，包含控制台、支持直接代码执行的语法高亮编辑器，以及用于绘图、历史记录、调试和工作区管理的工具。\n\n[Shiny](https:\u002F\u002Fshiny.rstudio.com\u002F) 是 RStudio 推出的一个较新的包，能够非常容易地使用 R 构建交互式 Web 应用程序。\n\n[Rmarkdown ](https:\u002F\u002Frmarkdown.rstudio.com\u002F) 是一个可以帮助你创建动态分析文档的包，这些文档可以将代码、渲染后的输出（如图表）和文字说明结合在一起。\n\n[Rplugin](https:\u002F\u002Fgithub.com\u002FJetBrains\u002FRplugin) 是 IntelliJ IDE 的 R 语言支持插件。\n\n[Plotly](https:\u002F\u002Fplotly-r.com\u002F) 是一个 R 包，可以通过开源 JavaScript 图形库 [plotly.js](https:\u002F\u002Fgithub.com\u002Fplotly\u002Fplotly.js) 创建交互式 Web 图形。\n\n[Metaflow](https:\u002F\u002Fmetaflow.org\u002F) 是一个 Python\u002FR 库，旨在帮助科学家和工程师构建并管理实际的数据科学项目。Metaflow 最初由 Netflix 开发，用于提高数据科学家的工作效率，他们需要处理从经典统计学到最先进深度学习等各种项目。\n\n[Prophet](https:\u002F\u002Ffacebook.github.io\u002Fprophet) 是一种基于加法模型的时间序列预测方法，其中非线性趋势通过年度、每周和每日季节性变化以及节假日效应来拟合。它最适合具有强烈季节性效应且有多个季节历史数据的时间序列。\n\n[LightGBM](https:\u002F\u002Flightgbm.readthedocs.io\u002F) 是一个基于树的学习算法的梯度提升框架，可用于排序、分类以及其他许多机器学习任务。\n\n[Dash](https:\u002F\u002Fplotly.com\u002Fdash) 是一个 Python 框架，可用于使用 Python、R、Julia 和 Jupyter 构建分析型 Web 应用程序。\n\n[MLR](https:\u002F\u002Fmlr.mlr-org.com\u002F) 是 R 中的机器学习工具。\n\n[ML workspace](https:\u002F\u002Fgithub.com\u002Fml-tooling\u002Fml-workspace) 是一款一体化的基于 Web 的集成开发环境，专为机器学习和数据科学而设计。它部署简单，几分钟内即可启动，帮助你在自己的机器上高效地构建机器学习解决方案。ML workspace 是开发者的终极工具，预装了多种流行的数据科学库（TensorFlow、PyTorch、Keras 和 MXNet）以及开发工具（Jupyter、VS Code 和 TensorBoard），所有这些都经过完美配置、优化和集成。\n\n[CatBoost](https:\u002F\u002Fcatboost.ai\u002F) 是一个快速、可扩展、高性能的决策树梯度提升库，适用于 Python、R、Java 和 C++，可用于排序、分类、回归及其他机器学习任务。它支持在 CPU 和 GPU 上进行计算。\n\n[Plumber](https:\u002F\u002Fwww.rplumber.io\u002F) 是一种工具，只需在现有的 R 源代码中添加特殊注释，即可创建 Web API。\n\n[Drake](https:\u002F\u002Fdocs.ropensci.org\u002Fdrake) 是一个专注于 R 的管道工具包，用于实现可重复性和高性能计算。\n\n[DiagrammeR](https:\u002F\u002Fvisualizers.co\u002Fdiagrammer\u002F) 是一个包，可以用来创建、修改、分析和可视化网络图。生成的图表可以嵌入到 R Markdown 文档中，集成到 Shiny Web 应用程序中，转换为其他图形格式，或导出为图像文件。\n\n[Knitr](https:\u002F\u002Fyihui.org\u002Fknitr\u002F) 是 R 中的一种通用文学编程引擎，提供轻量级的 API，使用户无需大量编码即可完全控制输出。\n\n[Broom](https:\u002F\u002Fbroom.tidymodels.org\u002F) 是一个将 R 中的统计分析对象转换为整洁格式的工具。\n\n# Julia 开发\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n\u003Cp align=\"center\">\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_readme_1424f02ea237.png\">\n  \u003Cbr \u002F>\n\u003C\u002Fp>\n\n# Julia 学习资源\n\n[Julia](https:\u002F\u002Fjulialang.org) 是一种高级、[高性能](https:\u002F\u002Fjulialang.org\u002Fbenchmarks\u002F) 的动态语言，适用于科学计算。Julia 程序通过 LLVM 编译为高效的本地代码，可在[多种平台](https:\u002F\u002Fjulialang.org\u002Fdownloads\u002F#support_tiers)上运行。\n\n[JuliaHub](https:\u002F\u002Fjuliahub.com\u002F) 包含超过 4,000 个供社区使用的 Julia 包。\n\n[Julia Observer](https:\u002F\u002Fwww.juliaobserver.com)\n\n[Julia 手册](https:\u002F\u002Fdocs.julialang.org\u002Fen\u002Fv1\u002Fmanual\u002Fgetting-started\u002F)\n\n[JuliaLang 基础知识](https:\u002F\u002Fdocs.julialang.org\u002Fen\u002Fv1\u002Fbase\u002Fbase\u002F)\n\n[Julia 风格指南](https:\u002F\u002Fdocs.julialang.org\u002Fen\u002Fv1\u002Fmanual\u002Fstyle-guide\u002F)\n\n[Julia 示例](https:\u002F\u002Fjuliabyexample.helpmanual.io\u002F)\n\n[JuliaLang Gitter](https:\u002F\u002Fgitter.im\u002FJuliaLang\u002Fjulia)\n\n[使用 Jupyter Notebooks 的 DataFrames 教程](https:\u002F\u002Fgithub.com\u002Fbkamins\u002FJulia-DataFrames-Tutorial\u002F)\n\n[Julia Academy](https:\u002F\u002Fjuliaacademy.com\u002Fcourses?preview=logged_out)\n\n[Julia 聚会小组](https:\u002F\u002Fwww.meetup.com\u002Ftopics\u002Fjulia\u002F)\n\n[Julia 在 Microsoft Azure 上](https:\u002F\u002Fjuliacomputing.com\u002Fmedia\u002F2017\u002F02\u002F08\u002Fazure.html)\n\n# Julia 工具\n\n[JuliaPro](https:\u002F\u002Fjuliacomputing.com\u002Fproducts\u002Fjuliapro.html) 是一种免费且快速的方式，适用于个人研究人员、工程师、科学家、量化分析师、交易员、经济学家、学生等。Julia 开发者可以更高效、更便捷地构建优质软件，同时受益于 Julia 无与伦比的高性能。它包含 2600 多个开源包，以及由 Julia Computing 精心挑选的 250 多个 JuliaPro 包。这些精选包经过测试、文档化，并由 Julia Computing 提供支持。\n\n[Juno](https:\u002F\u002Fjunolab.org) 是一款基于 [Atom]() 的功能强大的免费 IDE，专为 Julia 语言设计。\n\n[Debugger.jl](https:\u002F\u002Fgithub.com\u002FJuliaDebug\u002FDebugger.jl) 是 Julia 的调试工具。\n\n[Profile (Stdlib)](https:\u002F\u002Fdocs.julialang.org\u002Fen\u002Fv1\u002Fmanual\u002Fprofile\u002F) 是一个模块，提供帮助开发者优化代码性能的工具。使用时，它会对正在运行的代码进行度量，并生成输出，帮助你了解每行代码所花费的时间。\n\n[Revise.jl](https:\u002F\u002Fgithub.com\u002Ftimholy\u002FRevise.jl) 允许你在不重启 Julia 的情况下修改代码并立即生效。借助 Revise，你可以在会话中直接更新包、切换 Git 分支，或在任意编辑器中修改源代码；所有更改通常会在你从 REPL 发出的下一条命令时自动应用。这可以省去重启 Julia、加载包以及等待代码 JIT 编译的开销。\n\n[JuliaGPU](https:\u002F\u002Fjuliagpu.org\u002F) 是一个 GitHub 组织，旨在整合 Julia 中用于 GPU 编程的众多包。凭借其高级语法和灵活的编译器，Julia 在不牺牲性能的前提下，能够高效地对 GPU 等硬件加速器进行编程。\n\n[IJulia.jl](https:\u002F\u002Fgithub.com\u002FJuliaLang\u002FIJulia.jl) 是 Jupyter 的 Julia 内核。\n\n[AWS.jl](https:\u002F\u002Fgithub.com\u002FJuliaCloud\u002FAWS.jl) 是 Julia 对 [Amazon Web Services](https:\u002F\u002Faws.amazon.com\u002F) 的接口。\n\n[CUDA.jl](https:\u002F\u002Fjuliagpu.gitlab.io\u002FCUDA.jl) 是一个用于通过 Julia 操作 NVIDIA CUDA GPU 的主要编程接口包。它提供了易用的数组抽象、用于在 Julia 中编写 CUDA 核函数的编译器，以及对各种 CUDA 库的封装。\n\n[XLA.jl](https:\u002F\u002Fgithub.com\u002FJuliaTPU\u002FXLA.jl) 是一个将 Julia 编译为 XLA 的包，用于 [Tensor Processing Unit (TPU)](https:\u002F\u002Fcloud.google.com\u002Ftpu\u002F)。\n\n[Nanosoldier.jl](https:\u002F\u002Fgithub.com\u002FJuliaCI\u002FNanosoldier.jl) 是一个在 MIT 的 Nanosoldier 集群上运行 JuliaCI 服务的包。\n\n[Julia for VSCode](https:\u002F\u002Fwww.julia-vscode.org) 是 Julia 语言的强大扩展。\n\n[JuMP.jl](https:\u002F\u002Fjump.dev\u002F) 是一种嵌入在 Julia 中的领域特定建模语言，用于 [数学优化](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FMathematical_optimization)。\n\n[Optim.jl](https:\u002F\u002Fgithub.com\u002FJuliaNLSolvers\u002FOptim.jl) 是 Julia 中的一元和多元优化库。\n\n[RCall.jl](https:\u002F\u002Fgithub.com\u002FJuliaInterop\u002FRCall.jl) 是一个允许从 Julia 调用 R 函数的包。\n\n[JavaCall.jl](http:\u002F\u002Fjuliainterop.github.io\u002FJavaCall.jl) 是一个允许从 Julia 调用 Java 函数的包。\n\n[PyCall.jl](https:\u002F\u002Fgithub.com\u002FJuliaPy\u002FPyCall.jl) 是一个允许从 Julia 调用 Python 函数的包。\n\n[MXNet.jl](https:\u002F\u002Fgithub.com\u002Fdmlc\u002FMXNet.jl) 是 Apache MXNet 的 Julia 包。MXNet.jl 将灵活高效的 GPU 计算和最先进的深度学习引入 Julia。\n\n[Knet](https:\u002F\u002Fdenizyuret.github.io\u002FKnet.jl\u002Flatest) 是由 [Deniz Yuret](https:\u002F\u002Fwww.denizyuret.com\u002F) 及其合作者在 Julia 中实现的 [Koç University 深度](http:\u002F\u002Fwww.ku.edu.tr\u002Fen) 学习框架。它支持 GPU 运行和自动微分，并使用动态计算图来处理以纯 Julia 定义的模型。\n\n[Distributions.jl](https:\u002F\u002Fgithub.com\u002FJuliaStats\u002FDistributions.jl) 是一个用于概率分布及相关函数的 Julia 包。\n\n[DataFrames.jl](http:\u002F\u002Fjuliadata.github.io\u002FDataFrames.jl\u002Fstable\u002F) 是一个用于在 Julia 中处理表格数据的工具。\n\n[Flux.jl](https:\u002F\u002Ffluxml.ai\u002F) 是一种优雅的机器学习方法。它完全基于 Julia，提供了轻量级的抽象层，充分利用了 Julia 原生的 GPU 和自动微分支持。\n\n[IRTools.jl](https:\u002F\u002Fgithub.com\u002FFluxML\u002FIRTools.jl) 是一种简单而灵活的中间表示格式，足以处理降级后的 Julia 代码、类型化的 Julia 代码，以及外部的中间表示。\n\n[Cassette.jl](https:\u002F\u002Fgithub.com\u002Fjrevels\u002FCassette.jl) 是一个 Julia 包，它提供了一种机制，可以将代码转换阶段动态注入到 Julia 的即时编译（JIT）周期中，从而实现对“不了解 Cassette”的 Julia 程序的事后分析和修改，而无需手动添加源代码注释或重构目标代码。\n\n## 贡献\n\n- [x] 如果您希望为本指南做出贡献，请直接提交 [Pull Request](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide\u002Fpulls)。\n\n\n## 许可证\n\n[返回顶部](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#table-of-contents)\n\n根据 [知识共享署名 4.0 国际许可协议 (CC BY 4.0)](https:\u002F\u002Fcreativecommons.org\u002Flicenses\u002Fby\u002F4.0\u002F) 分发。","# Machine-Learning-Guide 快速上手指南\n\n**简介**：`Machine-Learning-Guide` 并非一个单一的代码库或可安装软件，而是一份由社区维护的**机器学习全景学习资源指南**。它汇集了从基础理论、经典教材、在线课程到主流框架（PyTorch, TensorFlow）、大模型（LLMs）工具链及行业最佳实践的权威链接。本指南旨在帮助开发者快速构建知识体系并找到合适的开发工具。\n\n## 1. 环境准备\n\n由于本指南主要提供资源索引，您无需为“指南”本身安装环境。但为了实践指南中推荐的工具（如 PyTorch, TensorFlow, Jupyter 等），建议准备以下开发环境：\n\n*   **操作系统**：Windows 10\u002F11, macOS (Intel\u002FApple Silicon), 或 Linux (Ubuntu\u002FCentOS 推荐)。\n*   **核心依赖**：\n    *   **Python**: 版本 3.8 - 3.11 (推荐通过 `conda` 或 `pyenv` 管理)。\n    *   **包管理器**: `pip` 或 `conda`。\n    *   **IDE\u002F编辑器**: VS Code (推荐安装 `Markdown PDF` 插件以便将本指南转为 PDF) 或 JupyterLab。\n*   **硬件加速 (可选但推荐)**：\n    *   NVIDIA GPU (需安装 CUDA Toolkit 和 cuDNN) 用于深度学习训练。\n    *   Apple Silicon (M1\u002FM2\u002FM3) 用户可利用 Core ML 或 MPS 加速。\n\n> **国内加速建议**：\n> 在安装 Python 依赖时，强烈建议使用国内镜像源以提升下载速度。\n> *   **清华源**: `https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple`\n> *   **阿里源**: `https:\u002F\u002Fmirrors.aliyun.com\u002Fpypi\u002Fsimple\u002F`\n\n## 2. 安装步骤\n\n您不需要“安装”此指南，而是需要**获取**它并根据您的需求安装指南中推荐的特定框架。\n\n### 第一步：获取指南内容\n克隆仓库或直接在浏览器阅读：\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide.git\ncd Machine-Learning-Guide\n```\n*或者直接在 GitHub 页面浏览目录结构。*\n\n### 第二步：搭建基础开发环境 (示例)\n根据指南中的 \"ML Frameworks\" 部分，以下是搭建通用机器学习环境的命令。\n\n**使用 Conda 创建隔离环境 (推荐):**\n```bash\n# 创建名为 ml_env 的环境，指定 Python 版本\nconda create -n ml_env python=3.9 -y\n\n# 激活环境\nconda activate ml_env\n\n# 配置国内镜像源 (以清华源为例)\nconda config --add channels https:\u002F\u002Fmirrors.tuna.tsinghua.edu.cn\u002Fanaconda\u002Fpkgs\u002Fmain\u002F\nconda config --add channels https:\u002F\u002Fmirrors.tuna.tsinghua.edu.cn\u002Fanaconda\u002Fpkgs\u002Ffree\u002F\nconda config --set show_channel_urls yes\n```\n\n**安装主流深度学习框架 (参考指南第 4、5 节):**\n```bash\n# 安装 PyTorch (CPU 版本，如需 GPU 请访问 pytorch.org 获取对应 CUDA 命令)\npip install torch torchvision torchaudio --index-url https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcpu\n\n# 或者安装 TensorFlow\npip install tensorflow\n\n# 安装数据处理与笔记本书写工具\npip install jupyterlab pandas scikit-learn matplotlib\n```\n\n## 3. 基本使用\n\n本指南的核心用法是**按需查阅**。以下是利用该指南进行学习和开发的标准流程：\n\n### 场景 A：寻找入门学习路径\n1.  打开仓库中的 `README.md`。\n2.  导航至 **[Learning Resources for ML](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#learning-resources-for-ml)** 章节。\n3.  **课程推荐**：点击 \"Courses & Certifications\"，选择 Andrew Ng 的 [Machine Learning by Stanford University](https:\u002F\u002Fwww.coursera.org\u002Flearn\u002Fmachine-learning) 开始理论基础学习。\n4.  **书籍推荐**：在 \"Books\" 章节下载《Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow》进行实战演练。\n\n### 场景 B：快速启动一个大模型 (LLM) 项目\n指南专门收录了 LLM 相关的工具链（见目录第 2 部分）。\n\n1.  查阅 **[Running Large Language Models (LLMs) Locally](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#running-llms-locally)** 章节。\n2.  根据指南推荐，选择适合本地的推理工具（如 `llama.cpp`, `Ollama`, 或 `LM Studio`）。\n3.  **示例：使用 pip 安装一个简单的 LLM 交互库 (基于指南推荐的生态)**：\n    ```bash\n    # 安装 transformers 库 (Hugging Face)\n    pip install transformers accelerate sentencepiece\n    \n    # 配置国内镜像加速 Hugging Face 模型下载\n    export HF_ENDPOINT=https:\u002F\u002Fhf-mirror.com\n    ```\n4.  编写简单的 Python 脚本测试：\n    ```python\n    from transformers import pipeline\n\n    # 加载一个轻量级的文本生成模型\n    generator = pipeline('text-generation', model='distilgpt2')\n\n    # 运行推理\n    result = generator(\"Hello, I am an AI developer learning from \", max_length=50)\n    print(result[0]['generated_text'])\n    ```\n\n### 场景 C：针对特定领域深入 (如计算机视觉或 NLP)\n1.  在目录中找到对应章节，例如 **[Computer Vision Development](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#computer-vision-development)** 或 **[NLP Development](https:\u002F\u002Fgithub.com\u002Fmikeroyal\u002FMachine-Learning-Guide#nlp-development)**。\n2.  获取该领域专用的库列表（如 OpenCV, YOLO, SpaCy, NLTK 等）。\n3.  参考指南中提供的 \"Developer Resources\" 链接，访问微软、Google 或 AWS 提供的该领域最佳实践代码库（Recipes\u002FCookbooks）。\n\n---\n*注：本指南内容会持续更新，建议定期 `git pull` 同步最新资源列表。*","某初创公司的算法工程师小李正负责从零搭建一个基于大语言模型（LLM）的智能客服系统，急需确定技术栈并寻找高质量的学习资源。\n\n### 没有 Machine-Learning-Guide 时\n- **选型迷茫**：面对 GitHub 上成千上万个框架和库，难以分辨哪些适合本地部署 LLM，哪些适合生产环境训练，浪费大量时间试错。\n- **资源分散**：寻找教程时需要同时在 Coursera、YouTube、技术博客和官方文档间跳转，缺乏系统性的学习路径指引。\n- **环境配置困难**：在配置 CUDA、PyTorch 或 TensorFlow 开发环境时，因缺少针对特定版本的避坑指南，频繁遭遇依赖冲突导致项目停滞。\n- **前沿滞后**：难以快速获取关于最新 LLM 训练框架和推理工具的一手信息，导致技术方案可能起步即落后。\n\n### 使用 Machine-Learning-Guide 后\n- **精准选型**：直接查阅\"LLMs Training Frameworks\"和\"Tools for deploying LLMs\"章节，快速锁定了适合当前算力条件的开源框架与部署工具。\n- **路径清晰**：利用\"Learning Resources\"板块中整理的课程、书籍及 YouTube 教程清单，为团队制定了从基础理论到实战开发的系统化培训计划。\n- **高效开发**：参考 PyTorch、TensorFlow 及 CUDA 开发专项指南，迅速解决了环境配置难题，将原本需要数天的调试工作压缩至几小时。\n- **技术领先**：通过持续更新的目录掌握最新的生物信息学、强化学习及 NLP 领域工具，确保产品架构始终处于行业前沿。\n\nMachine-Learning-Guide 将碎片化的机器学习生态整合为一张清晰的导航图，让开发者从“大海捞针”转变为“按图索骥”，极大提升了技术落地效率。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmikeroyal_Machine-Learning-Guide_57f321cb.png","mikeroyal","Michael Royal","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Fmikeroyal_652aad44.jpg","Senior Software Engineer by day.  A Linux and Open Source Contributor by night. 🤖",null,"California, USA","Miker256","https:\u002F\u002Fmikeroyal.github.io\u002F","https:\u002F\u002Fgithub.com\u002Fmikeroyal",[82],{"name":83,"color":84,"percentage":85},"Python","#3572A5",100,691,63,"2026-04-12T19:40:45",1,"","未说明",{"notes":93,"python":91,"dependencies":94},"该仓库并非一个可直接运行的软件工具，而是一份机器学习学习指南和资源列表。它主要包含课程、书籍、教程链接以及各类框架（如 PyTorch, TensorFlow, Core ML 等）的文档索引，因此没有具体的操作系统、硬件配置或依赖库安装要求。用户需根据指南中提到的具体子项目或框架去查阅相应的环境需求。",[],[35,15,14],[97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116],"python","deep-learning","machine-learning-library","machinelearning","machine-learning-models","pytorch","machinelearning-python","scikit-learn","scikitlearn-machine-learning","support-vector-machines","gpt-3","aws-sagemaker","artificial-neural-networks","jax","image-classification","gpt-4","generative-ai","image-processing","gpt4all","llms","2026-03-27T02:49:30.150509","2026-04-13T16:14:53.849183",[],[]]