[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-reczoo--FuxiCTR":3,"tool-reczoo--FuxiCTR":64},[4,17,27,35,43,56],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":16},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。",162132,3,"2026-04-05T11:01:52",[13,14,15],"开发框架","图像","Agent","ready",{"id":18,"name":19,"github_repo":20,"description_zh":21,"stars":22,"difficulty_score":23,"last_commit_at":24,"category_tags":25,"status":16},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",140436,2,"2026-04-05T23:32:43",[13,15,26],"语言模型",{"id":28,"name":29,"github_repo":30,"description_zh":31,"stars":32,"difficulty_score":23,"last_commit_at":33,"category_tags":34,"status":16},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。",107662,"2026-04-03T11:11:01",[13,14,15],{"id":36,"name":37,"github_repo":38,"description_zh":39,"stars":40,"difficulty_score":23,"last_commit_at":41,"category_tags":42,"status":16},3704,"NextChat","ChatGPTNextWeb\u002FNextChat","NextChat 是一款轻量且极速的 AI 助手，旨在为用户提供流畅、跨平台的大模型交互体验。它完美解决了用户在多设备间切换时难以保持对话连续性，以及面对众多 AI 模型不知如何统一管理的痛点。无论是日常办公、学习辅助还是创意激发，NextChat 都能让用户随时随地通过网页、iOS、Android、Windows、MacOS 或 Linux 端无缝接入智能服务。\n\n这款工具非常适合普通用户、学生、职场人士以及需要私有化部署的企业团队使用。对于开发者而言，它也提供了便捷的自托管方案，支持一键部署到 Vercel 或 Zeabur 等平台。\n\nNextChat 的核心亮点在于其广泛的模型兼容性，原生支持 Claude、DeepSeek、GPT-4 及 Gemini Pro 等主流大模型，让用户在一个界面即可自由切换不同 AI 能力。此外，它还率先支持 MCP（Model Context Protocol）协议，增强了上下文处理能力。针对企业用户，NextChat 提供专业版解决方案，具备品牌定制、细粒度权限控制、内部知识库整合及安全审计等功能，满足公司对数据隐私和个性化管理的高标准要求。",87618,"2026-04-05T07:20:52",[13,26],{"id":44,"name":45,"github_repo":46,"description_zh":47,"stars":48,"difficulty_score":23,"last_commit_at":49,"category_tags":50,"status":16},2268,"ML-For-Beginners","microsoft\u002FML-For-Beginners","ML-For-Beginners 是由微软推出的一套系统化机器学习入门课程，旨在帮助零基础用户轻松掌握经典机器学习知识。这套课程将学习路径规划为 12 周，包含 26 节精炼课程和 52 道配套测验，内容涵盖从基础概念到实际应用的完整流程，有效解决了初学者面对庞大知识体系时无从下手、缺乏结构化指导的痛点。\n\n无论是希望转型的开发者、需要补充算法背景的研究人员，还是对人工智能充满好奇的普通爱好者，都能从中受益。课程不仅提供了清晰的理论讲解，还强调动手实践，让用户在循序渐进中建立扎实的技能基础。其独特的亮点在于强大的多语言支持，通过自动化机制提供了包括简体中文在内的 50 多种语言版本，极大地降低了全球不同背景用户的学习门槛。此外，项目采用开源协作模式，社区活跃且内容持续更新，确保学习者能获取前沿且准确的技术资讯。如果你正寻找一条清晰、友好且专业的机器学习入门之路，ML-For-Beginners 将是理想的起点。",84991,"2026-04-05T10:45:23",[14,51,52,53,15,54,26,13,55],"数据工具","视频","插件","其他","音频",{"id":57,"name":58,"github_repo":59,"description_zh":60,"stars":61,"difficulty_score":10,"last_commit_at":62,"category_tags":63,"status":16},3128,"ragflow","infiniflow\u002Fragflow","RAGFlow 是一款领先的开源检索增强生成（RAG）引擎，旨在为大语言模型构建更精准、可靠的上下文层。它巧妙地将前沿的 RAG 技术与智能体（Agent）能力相结合，不仅支持从各类文档中高效提取知识，还能让模型基于这些知识进行逻辑推理和任务执行。\n\n在大模型应用中，幻觉问题和知识滞后是常见痛点。RAGFlow 通过深度解析复杂文档结构（如表格、图表及混合排版），显著提升了信息检索的准确度，从而有效减少模型“胡编乱造”的现象，确保回答既有据可依又具备时效性。其内置的智能体机制更进一步，使系统不仅能回答问题，还能自主规划步骤解决复杂问题。\n\n这款工具特别适合开发者、企业技术团队以及 AI 研究人员使用。无论是希望快速搭建私有知识库问答系统，还是致力于探索大模型在垂直领域落地的创新者，都能从中受益。RAGFlow 提供了可视化的工作流编排界面和灵活的 API 接口，既降低了非算法背景用户的上手门槛，也满足了专业开发者对系统深度定制的需求。作为基于 Apache 2.0 协议开源的项目，它正成为连接通用大模型与行业专有知识之间的重要桥梁。",77062,"2026-04-04T04:44:48",[15,14,13,26,54],{"id":65,"github_repo":66,"name":67,"description_en":68,"description_zh":69,"ai_summary_zh":69,"readme_en":70,"readme_zh":71,"quickstart_zh":72,"use_case_zh":73,"hero_image_url":74,"owner_login":75,"owner_name":76,"owner_avatar_url":77,"owner_bio":78,"owner_company":79,"owner_location":79,"owner_email":79,"owner_twitter":79,"owner_website":79,"owner_url":80,"languages":81,"stars":90,"forks":91,"last_commit_at":92,"license":93,"difficulty_score":23,"env_os":78,"env_gpu":94,"env_ram":95,"env_deps":96,"category_tags":102,"github_topics":103,"view_count":23,"oss_zip_url":79,"oss_zip_packed_at":79,"status":16,"created_at":109,"updated_at":110,"faqs":111,"releases":141},3823,"reczoo\u002FFuxiCTR","FuxiCTR","A configurable, tunable, and reproducible library for CTR prediction https:\u002F\u002Ffuxictr.github.io","FuxiCTR 是一个专为点击率（CTR）预测打造的开源算法库，广泛应用于在线广告、推荐系统及赞助搜索等工业场景。在学术界与工业界，CTR 模型的开发常面临代码复现难、参数调整繁琐以及不同框架间迁移成本高等痛点，FuxiCTR 正是为了解决这些问题而生。\n\n它主要服务于算法研究人员和工程开发者，帮助他们更高效地进行模型实验与落地。FuxiCTR 的核心优势在于其高度的可配置性、可调优性和可复现性。通过将数据预处理和模型结构模块化，用户只需通过简单的配置文件即可灵活组合实验流程，并支持自动化的超参数调优。此外，它同时兼容 PyTorch 和 TensorFlow 两大主流深度学习框架，极大地降低了新模型的开发与扩展门槛。\n\n目前，FuxiCTR 已集成了从经典的逻辑回归（LR）、因子分解机（FM）到深度语义匹配模型（DSSM）等多种经典与前沿的特征交互模型，并提供了标准的基准测试环境。无论是希望快速验证新想法的研究者，还是寻求稳定基线系统的工程师，都能利用 FuxiCTR 轻松复现论文结果或构建高性能的预测服务，从而推动相关领域的技术迭代与创新。","\u003Cdiv align=\"center\">\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Freczoo_FuxiCTR_readme_2478d4f99405.png\" alt=\"Logo\" width=\"260\"\u002F>\n\u003C\u002Fdiv>\n\n\u003Cdiv align=\"center\">\n\u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Ffuxictr\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpython-3.9+-blue\" style=\"max-width: 100%;\" alt=\"Python version\">\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Ffuxictr\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpytorch-1.10+-blue\" style=\"max-width: 100%;\" alt=\"Pytorch version\">\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Ffuxictr\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Ftensorflow-2.1+-blue\" style=\"max-width: 100%;\" alt=\"Pytorch version\">\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Ffuxictr\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Ffuxictr.svg\" style=\"max-width: 100%;\" alt=\"Pypi version\">\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fpepy.tech\u002Fproject\u002Ffuxictr\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Freczoo_FuxiCTR_readme_7ae11847d0dd.png\" style=\"max-width: 100%;\" alt=\"Downloads\">\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Freczoo\u002FFuxiCTR\u002Fblob\u002Fmain\u002FLICENSE\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flicense\u002Freczoo\u002Ffuxictr.svg\" style=\"max-width: 100%;\" alt=\"License\">\u003C\u002Fa>\n\u003C\u002Fdiv>\n\u003Chr\u002F>\n\n\u003Cdiv align=\"center\">\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Freczoo\u002FFuxiCTR\u002Fstargazers\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Freczoo_FuxiCTR_readme_f08b783e5dca.png\" width=\"600\"\u002F>\u003Ca\u002F>\n\u003C\u002Fdiv>\n\nClick-through rate (CTR) prediction is a critical task for various industrial applications such as online advertising, recommender systems, and sponsored search. FuxiCTR provides an open-source library for CTR prediction, with key features in configurability, tunability, and reproducibility. We hope this project could promote reproducible research and benefit both researchers and practitioners in this field.\n\n## Key Features\n\n+ **Configurable**: Both data preprocessing and models are modularized and configurable.\n\n+ **Tunable**: Models can be automatically tuned through easy configurations.\n\n+ **Reproducible**: All the benchmarks can be easily reproduced.\n\n+ **Extensible**: It can be easily extended to any new models, supporting both Pytorch and Tensorflow frameworks.\n\n\n## Model Zoo\n\n| No  | Publication       | Model                                    | Paper                                                                                                                                                                                                           | Benchmark                                                                                                       | Version       |\n|:---:|:-----------------:|:----------------------------------------:|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |:---------------------------------------------------------------------------------------------------------------:|:-------------:|\n|\u003Ctr>\u003Cth colspan=6 align=\"center\">:open_file_folder: **Feature Interaction Models**\u003C\u002Fth>\u003C\u002Ftr>|\n| 1   | WWW'07            | [LR](.\u002Fmodel_zoo\u002FLR)                     | [Predicting Clicks: Estimating the Click-Through Rate for New Ads](https:\u002F\u002Fdl.acm.org\u002Fcitation.cfm?id=1242643) :triangular_flag_on_post:**Microsoft**                                                           | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FLR)           | `torch`       |\n| 2   | ICDM'10           | [FM](.\u002Fmodel_zoo\u002FFM)                     | [Factorization Machines](https:\u002F\u002Fwww.csie.ntu.edu.tw\u002F~b97053\u002Fpaper\u002FRendle2010FM.pdf)                                                                                                                            | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFM)           | `torch`       |\n| 3   | CIKM'13           | [DSSM](.\u002Fmodel_zoo\u002FDSSM)                 | [Learning Deep Structured Semantic Models  for Web Search using Clickthrough Data ](https:\u002F\u002Fposenhuang.github.io\u002Fpapers\u002Fcikm2013_DSSM_fullversion.pdf) :triangular_flag_on_post:**Microsoft**                   | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDSSM)         | `torch`       |\n| 4   | CIKM'15           | [CCPM](.\u002Fmodel_zoo\u002FCCPM)                 | [A Convolutional Click Prediction Model](http:\u002F\u002Fwww.escience.cn\u002Fsystem\u002Fdownload\u002F73676)                                                                                                                          | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FCCPM)         | `torch`       |\n| 5   | RecSys'16         | [FFM](.\u002Fmodel_zoo\u002FFFM)                   | [Field-aware Factorization Machines for CTR Prediction](https:\u002F\u002Fdl.acm.org\u002Fcitation.cfm?id=2959134) :triangular_flag_on_post:**Criteo**                                                                         | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFFM)          | `torch`       |\n| 6   | RecSys'16         | [DNN](.\u002Fmodel_zoo\u002FDNN)            | [Deep Neural Networks for YouTube Recommendations](http:\u002F\u002Fart.yale.edu\u002Ffile_columns\u002F0001\u002F1132\u002Fcovington.pdf) :triangular_flag_on_post:**Google**                                                                | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDNN)          | `torch`, `tf` |\n| 7   | DLRS'16           | [Wide&Deep](.\u002Fmodel_zoo\u002FWideDeep)        | [Wide & Deep Learning for Recommender Systems](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1606.07792.pdf) :triangular_flag_on_post:**Google**                                                                                        | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FWideDeep)     | `torch`, `tf` |\n| 8   | ICDM'16           | [PNN](.\u002Fmodel_zoo\u002FPNN)                  | [Product-based Neural Networks for User Response Prediction](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1611.00144.pdf)                                                                                                              | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FPNN)         | `torch`       |\n| 9   | KDD'16            | [DeepCrossing](.\u002Fmodel_zoo\u002FDeepCrossing) | [Deep Crossing: Web-Scale Modeling without Manually Crafted Combinatorial Features](https:\u002F\u002Fwww.kdd.org\u002Fkdd2016\u002Fpapers\u002Ffiles\u002Fadf0975-shanA.pdf) :triangular_flag_on_post:**Microsoft**                          | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDeepCrossing) | `torch`       |\n| 10  | NIPS'16           | [HOFM](.\u002Fmodel_zoo\u002FHOFM)                 | [Higher-Order Factorization Machines](https:\u002F\u002Fpapers.nips.cc\u002Fpaper\u002F6144-higher-order-factorization-machines.pdf)                                                                                                | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FHOFM)         | `torch`       |\n| 11  | IJCAI'17          | [DeepFM](.\u002Fmodel_zoo\u002FDeepFM)             | [DeepFM: A Factorization-Machine based Neural Network for CTR Prediction](https:\u002F\u002Farxiv.org\u002Fabs\u002F1703.04247) :triangular_flag_on_post:**Huawei**                                                                 | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDeepFM)       | `torch`, `tf` |\n| 12  | SIGIR'17          | [NFM](.\u002Fmodel_zoo\u002FNFM)                   | [Neural Factorization Machines for Sparse Predictive Analytics](https:\u002F\u002Fdl.acm.org\u002Fcitation.cfm?id=3080777)                                                                                                     | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FNFM)          | `torch`       |\n| 13  | IJCAI'17          | [AFM](.\u002Fmodel_zoo\u002FAFM)                   | [Attentional Factorization Machines: Learning the Weight of Feature Interactions via Attention Networks](http:\u002F\u002Fwww.ijcai.org\u002Fproceedings\u002F2017\u002F0435.pdf)                                                        | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FAFM)          | `torch`       |\n| 14  | ADKDD'17          | [DCN](.\u002Fmodel_zoo\u002FDCN)                   | [Deep & Cross Network for Ad Click Predictions](https:\u002F\u002Farxiv.org\u002Fabs\u002F1708.05123) :triangular_flag_on_post:**Google**                                                                                           | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDCN)          | `torch`, `tf` |\n| 15  | WWW'18            | [FwFM](.\u002Fmodel_zoo\u002FFwFM)                 | [Field-weighted Factorization Machines for Click-Through Rate Prediction in Display Advertising](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1806.03514.pdf) :triangular_flag_on_post:**Oath, TouchPal, LinkedIn, Alibaba**           | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFwFM)         | `torch`       |\n| 16  | KDD'18            | [xDeepFM](.\u002Fmodel_zoo\u002FxDeepFM)           | [xDeepFM: Combining Explicit and Implicit Feature Interactions for Recommender Systems](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1803.05170.pdf) :triangular_flag_on_post:**Microsoft**                                            | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FxDeepFM)      | `torch`       |\n| 17  | CIKM'19           | [FiGNN](.\u002Fmodel_zoo\u002FFiGNN)               | [FiGNN: Modeling Feature Interactions via Graph Neural Networks for CTR Prediction](https:\u002F\u002Farxiv.org\u002Fabs\u002F1910.05552)                                                                                           | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFiGNN)        | `torch`       |\n| 18  | CIKM'19           | [AutoInt\u002FAutoInt+](.\u002Fmodel_zoo\u002FAutoInt)  | [AutoInt: Automatic Feature Interaction Learning via Self-Attentive Neural Networks](https:\u002F\u002Farxiv.org\u002Fabs\u002F1810.11921)                                                                                          | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FAutoInt)      | `torch`       |\n| 19  | RecSys'19         | [FiBiNET](.\u002Fmodel_zoo\u002FFiBiNET)           | [FiBiNET: Combining Feature Importance and Bilinear feature Interaction for Click-Through Rate Prediction](https:\u002F\u002Farxiv.org\u002Fabs\u002F1905.09433) :triangular_flag_on_post:**Sina Weibo**                            | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFiBiNET)      | `torch`       |\n| 20  | WWW'19            | [FGCNN](.\u002Fmodel_zoo\u002FFGCNN)               | [Feature Generation by Convolutional Neural Network for Click-Through Rate Prediction](https:\u002F\u002Farxiv.org\u002Fabs\u002F1904.04447) :triangular_flag_on_post:**Huawei**                                                    | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFGCNN)        | `torch`       |\n| 21  | AAAI'19           | [HFM\u002FHFM+](.\u002Fmodel_zoo\u002FHFM)              | [Holographic Factorization Machines for Recommendation](https:\u002F\u002Fojs.aaai.org\u002F\u002Findex.php\u002FAAAI\u002Farticle\u002Fview\u002F4448)                                                                                                 | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FHFM)          | `torch`       |\n| 22  | Arxiv'19          | [DLRM](.\u002Fmodel_zoo\u002FDLRM)                 | [Deep Learning Recommendation Model for Personalization and Recommendation Systems](https:\u002F\u002Farxiv.org\u002Fabs\u002F1906.00091) :triangular_flag_on_post:**Facebook**                                                     | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDLRM)         | `torch`       |\n| 23  | NeuralNetworks'20 | [ONN](.\u002Fmodel_zoo\u002FONN)                   | [Operation-aware Neural Networks for User Response Prediction](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1904.12579)                                                                                                                | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FONN)          | `torch`, `tf`      |\n| 24  | AAAI'20           | [AFN\u002FAFN+](.\u002Fmodel_zoo\u002FAFN)              | [Adaptive Factorization Network: Learning Adaptive-Order Feature Interactions](https:\u002F\u002Fojs.aaai.org\u002Findex.php\u002FAAAI\u002Farticle\u002Fview\u002F5768)                                                                           | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FAFN)          | `torch`       |\n| 25  | AAAI'20           | [LorentzFM](.\u002Fmodel_zoo\u002FLorentzFM)       | [Learning Feature Interactions with Lorentzian Factorization](https:\u002F\u002Farxiv.org\u002Fabs\u002F1911.09821) :triangular_flag_on_post:**eBay**                                                                               | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FLorentzFM)    | `torch`       |\n| 26  | WSDM'20           | [InterHAt](.\u002Fmodel_zoo\u002FInterHAt)         | [Interpretable Click-through Rate Prediction through Hierarchical Attention](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002F10.1145\u002F3336191.3371785) :triangular_flag_on_post:**NEC Labs, Google**                                      | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FInterHAt)     | `torch`       |\n| 27  | DLP-KDD'20        | [FLEN](.\u002Fmodel_zoo\u002FFLEN)                 | [FLEN: Leveraging Field for Scalable CTR Prediction](https:\u002F\u002Farxiv.org\u002Fabs\u002F1911.04690) :triangular_flag_on_post:**Tencent**                                                                                     | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFLEN)         | `torch`       |\n| 28  | CIKM'20           | [DeepIM](.\u002Fmodel_zoo\u002FDeepIM)             | [Deep Interaction Machine: A Simple but Effective Model for High-order Feature Interactions](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002Fabs\u002F10.1145\u002F3340531.3412077) :triangular_flag_on_post:**Alibaba, RealAI**                   | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDeepIM)       | `torch`       |\n| 29  | WWW'21            | [FmFM](.\u002Fmodel_zoo\u002FFmFM)                 | [FM^2: Field-matrixed Factorization Machines for Recommender Systems](https:\u002F\u002Farxiv.org\u002Fabs\u002F2102.12994) :triangular_flag_on_post:**Yahoo**                                                                      | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFmFM)         | `torch`       |\n| 30  | WWW'21            | [DCN-V2](.\u002Fmodel_zoo\u002FDCNv2)              | [DCN V2: Improved Deep & Cross Network and Practical Lessons for Web-scale Learning to Rank Systems](https:\u002F\u002Farxiv.org\u002Fabs\u002F2008.13535) :triangular_flag_on_post:**Google**                                      | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDCNv2)        | `torch`       |\n| 31  | CIKM'21           | [DESTINE](.\u002Fmodel_zoo\u002FDESTINE)           | [Disentangled Self-Attentive Neural Networks for Click-Through Rate Prediction](https:\u002F\u002Farxiv.org\u002Fabs\u002F2101.03654) :triangular_flag_on_post:**Alibaba**                                                          | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDESTINE)      | `torch`       |\n| 32  | CIKM'21           | [EDCN](.\u002Fmodel_zoo\u002FEDCN)                 | [Enhancing Explicit and Implicit Feature Interactions via Information Sharing for Parallel Deep CTR Models](https:\u002F\u002Fdlp-kdd.github.io\u002Fassets\u002Fpdf\u002FDLP-KDD_2021_paper_12.pdf) :triangular_flag_on_post:**Huawei** | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FEDCN)         | `torch`       |\n| 33  | DLP-KDD'21        | [MaskNet](.\u002Fmodel_zoo\u002FMaskNet)           | [MaskNet: Introducing Feature-Wise Multiplication to CTR Ranking Models by Instance-Guided Mask](https:\u002F\u002Farxiv.org\u002Fabs\u002F2102.07619) :triangular_flag_on_post:**Sina Weibo**                                      | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FMaskNet)      | `torch`       |\n| 34  | SIGIR'21          | [SAM](.\u002Fmodel_zoo\u002FSAM)                   | [Looking at CTR Prediction Again: Is Attention All You Need?](https:\u002F\u002Farxiv.org\u002Fabs\u002F2105.05563) :triangular_flag_on_post:**BOSS Zhipin**                                                                        | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FSAM)          | `torch`       |\n| 35  | KDD'21            | [AOANet](.\u002Fmodel_zoo\u002FAOANet)             | [Architecture and Operation Adaptive Network for Online Recommendations](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002F10.1145\u002F3447548.3467133) :triangular_flag_on_post:**Didi Chuxing**                                              | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FAOANet)       | `torch`       |\n| 36  | AAAI'23           | [FinalMLP](.\u002Fmodel_zoo\u002FFinalMLP)         | [FinalMLP: An Enhanced Two-Stream MLP Model for CTR Prediction](https:\u002F\u002Farxiv.org\u002Fabs\u002F2304.00902) :triangular_flag_on_post:**Huawei**                                                                                                               |     [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFinalMLP)         | `torch`       |\n| 37  | SIGIR'23          | [FinalNet](.\u002Fmodel_zoo\u002FFinalNet)               | [FINAL: Factorized Interaction Layer for CTR Prediction](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002F10.1145\u002F3539618.3591988) :triangular_flag_on_post:**Huawei**                                                                                                               |     [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFinalNet)         | `torch`       |\n| 38  | SIGIR'23          | [EulerNet](.\u002Fmodel_zoo\u002FEulerNet)               | [EulerNet: Adaptive Feature Interaction Learning via Euler's Formula for CTR Prediction](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002F10.1145\u002F3539618.3591681) :triangular_flag_on_post:**Huawei**                                                                                                               |     [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002FEthan-TZ\u002FEulerNet\u002Ftree\u002Fmain\u002F%23Code4FuxiCTR%23)         | `torch`       |\n| 39  | CIKM'23           | [GDCN](.\u002Fmodel_zoo\u002FGDCN)         | [Towards Deeper, Lighter and Interpretable Cross Network for CTR Prediction](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002Fpdf\u002F10.1145\u002F3583780.3615089) :triangular_flag_on_post:**Microsoft**                                                                                                               |           | `torch`       |\n| 40  | ICML'24          | [WuKong](.\u002Fmodel_zoo\u002FWuKong)               | [Wukong: Towards a Scaling Law for Large-Scale Recommendation](https:\u002F\u002Farxiv.org\u002Fabs\u002F2403.02545) :triangular_flag_on_post:**Meta**                                                        |   [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FWuKong)    | `torch`       |\n|\u003Ctr>\u003Cth colspan=6 align=\"center\">:open_file_folder: **Behavior Sequence Modeling**\u003C\u002Fth>\u003C\u002Ftr>|\n| 42  | KDD'18            | [DIN](.\u002Fmodel_zoo\u002FDIN)                   | [Deep Interest Network for Click-Through Rate Prediction](https:\u002F\u002Fwww.kdd.org\u002Fkdd2018\u002Faccepted-papers\u002Fview\u002Fdeep-interest-network-for-click-through-rate-prediction) :triangular_flag_on_post:**Alibaba**        |   [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDIN)       | `torch`       |\n| 43  | AAAI'19           | [DIEN](.\u002Fmodel_zoo\u002FDIEN)                 | [Deep Interest Evolution Network for Click-Through Rate Prediction](https:\u002F\u002Farxiv.org\u002Fabs\u002F1809.03672) :triangular_flag_on_post:**Alibaba**                                                                      |   [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDIEN)        | `torch`       |\n| 44  | DLP-KDD'19        | [BST](.\u002Fmodel_zoo\u002FBST)                   | [Behavior Sequence Transformer for E-commerce Recommendation in Alibaba](https:\u002F\u002Farxiv.org\u002Fabs\u002F1905.06874) :triangular_flag_on_post:**Alibaba**                                                                 |  [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FBST)     | `torch`       |\n| 45  | CIKM'20           | [DMIN](.\u002Fmodel_zoo\u002FDMIN)                 | [Deep Multi-Interest Network for Click-through Rate Prediction](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002F10.1145\u002F3340531.3412092) :triangular_flag_on_post:**Alibaba**                                                            | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDMIN)                                                                                                                 | `torch`       |\n| 46  | AAAI'20           | [DMR](.\u002Fmodel_zoo\u002FDMR)                   | [Deep Match to Rank Model for Personalized Click-Through Rate Prediction](https:\u002F\u002Fojs.aaai.org\u002Findex.php\u002FAAAI\u002Farticle\u002Fview\u002F5346) :triangular_flag_on_post:**Alibaba**                                           |    [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDMR)                                                                                                                  | `torch`       |\n| 47  | KDD'23           | [TransAct](.\u002Fmodel_zoo\u002FTransAct)                 | [TransAct: Transformer-based Realtime User Action Model for Recommendation at Pinterest](https:\u002F\u002Farxiv.org\u002Fabs\u002F2306.00248) :triangular_flag_on_post:**Pinterest**                                                       | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FTransAct)         | `torch`       |\n|\u003Ctr>\u003Cth colspan=6 align=\"center\">:open_file_folder: **Long Sequence Modeling**\u003C\u002Fth>\u003C\u002Ftr>|\n| 48  | CIKM'20          | [SIM](.\u002Fmodel_zoo\u002FLongCTR\u002FSIM)                   | [Search-based User Interest Modeling with Lifelong Sequential Behavior Data for Click-Through Rate Prediction](https:\u002F\u002Farxiv.org\u002Fabs\u002F2006.05639) :triangular_flag_on_post:**Alibaba**                                                               |                                                                                                                 | `torch`       |\n| 49  | DLP-KDD'22          | [ETA](.\u002Fmodel_zoo\u002FLongCTR\u002FETA)                   | [Efficient Long Sequential User Data Modeling for Click-Through Rate Prediction](https:\u002F\u002Farxiv.org\u002Fabs\u002F2209.12212) :triangular_flag_on_post:**Alibaba**                                                               |                                                                                                                 | `torch`       |\n| 50  | CIKM'22           | [SDIM](.\u002Fmodel_zoo\u002FLongCTR\u002FSDIM)                 | [Sampling Is All You Need on Modeling Long-Term User Behaviors for CTR Prediction](https:\u002F\u002Farxiv.org\u002Fabs\u002F2205.10249) :triangular_flag_on_post:**Meituan**                                                       |                                                                                                                 | `torch`       |\n| 51  | KDD'23           | [TWIN](.\u002Fmodel_zoo\u002FLongCTR\u002FTWIN)                 | [TWIN: TWo-stage Interest Network for Lifelong User Behavior Modeling in CTR Prediction at Kuaishou](https:\u002F\u002Farxiv.org\u002Fabs\u002F2302.02352) :triangular_flag_on_post:**KuaiShou**                                                       |                                                                                                                 | `torch`       |\n| 52  | KDD'25           | [MIRRN](.\u002Fmodel_zoo\u002FLongCTR\u002FMIRRN)                 | [Multi-granularity Interest Retrieval and Refinement Network for Long-Term User Behavior Modeling in CTR Prediction](https:\u002F\u002Farxiv.org\u002Fabs\u002F2411.15005) :triangular_flag_on_post:**Huawei**                                                       |                                                                                                                 | `torch`       |\n|\u003Ctr>\u003Cth colspan=6 align=\"center\">:open_file_folder: **Dynamic Weight Network**\u003C\u002Fth>\u003C\u002Ftr>|\n| 53  | NeurIPS'22          | [APG](.\u002Fmodel_zoo\u002FAPG)               | [APG: Adaptive Parameter Generation Network for Click-Through Rate Prediction](https:\u002F\u002Farxiv.org\u002Fabs\u002F2203.16218) :triangular_flag_on_post:**Alibaba**                                |    [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FAPG)                                                                                                   | `torch`       |\n| 54  | KDD'23        | [PPNet](.\u002Fmodel_zoo\u002FPEPNet)             | [PEPNet: Parameter and Embedding Personalized Network for Infusing with Personalized Prior Information](https:\u002F\u002Farxiv.org\u002Fabs\u002F2302.01115) :triangular_flag_on_post:**KuaiShou**                          |    [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FPPNet)                                                                                                   | `torch`       |\n|\u003Ctr>\u003Cth colspan=6 align=\"center\">:open_file_folder: **Multi-Task Modeling**\u003C\u002Fth>\u003C\u002Ftr>|\n| 55  |     Arxiv'17      | [ShareBottom](.\u002Fmodel_zoo\u002Fmultitask\u002FShareBottom)               | [An Overview of Multi-Task Learning in Deep Neural Networks](https:\u002F\u002Farxiv.org\u002Fabs\u002F1706.05098)                                                                                            |                                                                                                                 | `torch`       |\n| 56  | KDD'18          | [MMoE](.\u002Fmodel_zoo\u002Fmultitask\u002FMMOE)               | [Modeling Task Relationships in Multi-task Learning with Multi-Gate Mixture-of-Experts](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002Fpdf\u002F10.1145\u002F3219819.3220007) :triangular_flag_on_post:**Google**                                                                                            |                                                                                                                 | `torch`       |\n| 57  | RecSys'20          | [PLE](.\u002Fmodel_zoo\u002Fmultitask\u002FPLE)               | [Progressive Layered Extraction (PLE): A Novel Multi-Task Learning (MTL) Model for Personalized Recommendations](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002F10.1145\u002F3383313.3412236) :triangular_flag_on_post:**Tencent**                                                                                            |                                                                                                                 | `torch`       |\n\n## Benchmarking\n\nWe have benchmarked FuxiCTR models on a set of open datasets as follows:\n\n+ :star: [Benchmark datasets for CTR prediction](https:\u002F\u002Fgithub.com\u002Freczoo\u002FDatasets?tab=readme-ov-file#ctr-prediction)\n+ :star: [Benchmark settings and running steps](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr)\n+ :star: [Benchmark leaderboard for CTR prediction](https:\u002F\u002Fopenbenchmark.github.io\u002FBARS\u002FCTR\u002Fleaderboard)\n\n## Dependencies\n\nFuxiCTR has the following dependencies:\n\n+ python 3.9+\n+ pytorch 1.10.0--2.1.2 (if using for torch models)\n+ tensorflow 2.1 (if using for tensorflow models)\n\nPlease install other required packages via `pip install -r requirements.txt`.\n\n## Quick Start\n\n1. Run the demo examples\n   \n    Examples are provided in the demo directory to show some basic usage of FuxiCTR. Users can run the examples for quick start and to understand the workflow. \n   \n   ```\n   cd demo\n   python example1_build_dataset_to_parquet.py\n   python example2_DeepFM_with_parquet_input.py\n   ```\n\n2. Run a model on tiny data\n   \n    Users can easily run each model in the model zoo following the commands below, which is a demo for running DCN. In addition, users can modify the dataset config and model config files to run on their own datasets or with new hyper-parameters. More details can be found in the [README](.\u002Fmodel_zoo\u002FDCN\u002FDCN_torch\u002FREADME.md).\n   \n   ```\n   cd model_zoo\u002FDCN\u002FDCN_torch\n   python run_expid.py --expid DCN_test --gpu 0\n\n   # Change `MODEL` according to the target model name\n   cd model_zoo\u002FMODEL\n   python run_expid.py --expid MODEL_test --gpu 0\n   ```\n\n3. Run a model on benchmark datasets (e.g., Criteo)\n\n   Users can follow the [benchmark section](#Benchmarking) to get benchmark datasets and running steps for reproducing the existing results. Please see an example here: https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDCNv2\u002FDCNv2_criteo_x1\n\n\n4. Implement a new model\n   \n   The FuxiCTR library is designed to be modularized, so that every component can be overwritten by users according to their needs. In many cases, only the model class needs to be implemented for a new customized model. If data preprocessing or data loader is not directly applicable, one can also overwrite a new one through the [core APIs](https:\u002F\u002Fwww.processon.com\u002Fview\u002Flink\u002F63cfcfab4e30670eac4a81c7). We show a concrete example which implements our new model [FinalMLP](https:\u002F\u002Freczoo.github.io\u002FFinalMLP) that has been recently published in AAAI 2023.\n\n5. Tune hyper-parameters of a model\n   \n   FuxiCTR currently support fast grid search of hyper-parameters of a model using multiple GPUs. The following example shows the grid search of 8 experiments with 4 GPUs.\n    \n   ```\n   cd experiment\n   python run_param_tuner.py --config config\u002FDCN_tiny_parquet_tuner_config.yaml --gpu 0 1 2 3 0 1 2 3\n   ```\n\n## 🔥 Citation\n\nIf you use our code or benchmarks in your public research, please cite the following two papers.\n\n+ Jieming Zhu, Quanyu Dai, Liangcai Su, Rong Ma, Jinyang Liu, Guohao Cai, Xi Xiao, Rui Zhang. [BARS: Towards Open Benchmarking for Recommender Systems](https:\u002F\u002Farxiv.org\u002Fabs\u002F2205.09626). *The 45th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR)*, 2022. [[Bibtex](https:\u002F\u002Fdblp.org\u002Frec\u002Fconf\u002Fsigir\u002FZhuDSMLCXZ22.html?view=bibtex)]\n+ Jieming Zhu, Jinyang Liu, Shuai Yang, Qi Zhang, Xiuqiang He. [BARS-CTR: Open Benchmarking for Click-Through Rate Prediction](https:\u002F\u002Farxiv.org\u002Fabs\u002F2009.05794). *The 30th ACM International Conference on Information and Knowledge Management (CIKM)*, 2021. [[Bibtex](https:\u002F\u002Fdblp.org\u002Frec\u002Fconf\u002Fcikm\u002FZhuLYZH21.html?view=bibtex)]\n\n## 🙋 Discussion\n\nWelcome to join our WeChat group for any question and discussion. If you are interested in research and practice in recommender systems, please reach out via our WeChat group.\n\n![Scan QR code](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Freczoo_FuxiCTR_readme_b19ca98f40a7.jpg)\n","\u003Cdiv align=\"center\">\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Freczoo_FuxiCTR_readme_2478d4f99405.png\" alt=\"Logo\" width=\"260\"\u002F>\n\u003C\u002Fdiv>\n\n\u003Cdiv align=\"center\">\n\u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Ffuxictr\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpython-3.9+-blue\" style=\"max-width: 100%;\" alt=\"Python版本\">\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Ffuxictr\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpytorch-1.10+-blue\" style=\"max-width: 100%;\" alt=\"PyTorch版本\">\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Ffuxictr\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Ftensorflow-2.1+-blue\" style=\"max-width: 100%;\" alt=\"TensorFlow版本\">\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Ffuxictr\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Ffuxictr.svg\" style=\"max-width: 100%;\" alt=\"PyPI版本\">\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fpepy.tech\u002Fproject\u002Ffuxictr\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Freczoo_FuxiCTR_readme_7ae11847d0dd.png\" style=\"max-width: 100%;\" alt=\"下载量\">\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Freczoo\u002FFuxiCTR\u002Fblob\u002Fmain\u002FLICENSE\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flicense\u002Freczoo\u002Ffuxictr.svg\" style=\"max-width: 100%;\" alt=\"许可证\">\u003C\u002Fa>\n\u003C\u002Fdiv>\n\u003Chr\u002F>\n\n\u003Cdiv align=\"center\">\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Freczoo\u002FFuxiCTR\u002Fstargazers\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Freczoo_FuxiCTR_readme_f08b783e5dca.png\" width=\"600\"\u002F>\u003Ca\u002F>\n\u003C\u002Fdiv>\n\n点击率（CTR）预测是在线广告、推荐系统和赞助搜索等多种工业应用中的关键任务。FuxiCTR 提供了一个用于 CTR 预测的开源库，其主要特点在于可配置性、可调性和可复现性。我们希望该项目能够促进可复现性研究，并使该领域的研究人员和从业者受益。\n\n## 主要特性\n\n+ **可配置**：数据预处理和模型均采用模块化设计，且高度可配置。\n\n+ **可调优**：通过简单的配置即可实现模型的自动调优。\n\n+ **可复现**：所有基准测试均可轻松复现。\n\n+ **可扩展**：可以方便地扩展到任何新模型，同时支持 PyTorch 和 TensorFlow 框架。\n\n\n## 模型库\n\n| 序号 | 出版物       | 模型                                    | 论文                                                                                                                                                                                                           | 基准                                                                                                       | 版本       |\n|:---:|:-----------------:|:----------------------------------------:|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |:---------------------------------------------------------------------------------------------------------------:|:-------------:|\n|\u003Ctr>\u003Cth colspan=6 align=\"center\">:open_file_folder: **特征交互模型**\u003C\u002Fth>\u003C\u002Ftr>|\n| 1   | WWW'07            | [LR](.\u002Fmodel_zoo\u002FLR)                     | [预测点击：为新广告估计点击率](https:\u002F\u002Fdl.acm.org\u002Fcitation.cfm?id=1242643) :triangular_flag_on_post:**微软**                                                           | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FLR)           | `torch`       |\n| 2   | ICDM'10           | [FM](.\u002Fmodel_zoo\u002FFM)                     | [因子分解机](https:\u002F\u002Fwww.csie.ntu.edu.tw\u002F~b97053\u002Fpaper\u002FRendle2010FM.pdf)                                                                                                                            | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFM)           | `torch`       |\n| 3   | CIKM'13           | [DSSM](.\u002Fmodel_zoo\u002FDSSM)                 | [使用点击数据学习用于网络搜索的深度结构化语义模型](https:\u002F\u002Fposenhuang.github.io\u002Fpapers\u002Fcikm2013_DSSM_fullversion.pdf) :triangular_flag_on_post:**微软**                   | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDSSM)         | `torch`       |\n| 4   | CIKM'15           | [CCPM](.\u002Fmodel_zoo\u002FCCPM)                 | [卷积点击预测模型](http:\u002F\u002Fwww.escience.cn\u002Fsystem\u002Fdownload\u002F73676)                                                                                                                          | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FCCPM)         | `torch`       |\n| 5   | RecSys'16         | [FFM](.\u002Fmodel_zoo\u002FFFM)                   | [用于点击率预测的领域感知因子分解机](https:\u002F\u002Fdl.acm.org\u002Fcitation.cfm?id=2959134) :triangular_flag_on_post:**Criteo**                                                                         | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFFM)          | `torch`       |\n| 6   | RecSys'16         | [DNN](.\u002Fmodel_zoo\u002FDNN)            | [YouTube 推荐系统的深度神经网络](http:\u002F\u002Fart.yale.edu\u002Ffile_columns\u002F0001\u002F1132\u002Fcovington.pdf) :triangular_flag_on_post:**谷歌**                                                                | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDNN)          | `torch`, `tf` |\n| 7   | DLRS'16           | [Wide&Deep](.\u002Fmodel_zoo\u002FWideDeep)        | [推荐系统中的 Wide & Deep 学习](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1606.07792.pdf) :triangular_flag_on_post:**谷歌**                                                                                        | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FWideDeep)     | `torch`, `tf` |\n| 8   | ICDM'16           | [PNN](.\u002Fmodel_zoo\u002FPNN)                  | [基于产品的神经网络用于用户响应预测](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1611.00144.pdf)                                                                                                              | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FPNN)         | `torch`       |\n| 9   | KDD'16            | [DeepCrossing](.\u002Fmodel_zoo\u002FDeepCrossing) | [Deep Crossing：无需手工设计组合特征的Web-scale建模](https:\u002F\u002Fwww.kdd.org\u002Fkdd2016\u002Fpapers\u002Ffiles\u002Fadf0975-shanA.pdf) :triangular_flag_on_post:**微软**                          | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDeepCrossing) | `torch`       |\n| 10  | NIPS'16           | [HOFM](.\u002Fmodel_zoo\u002FHOFM)                 | [高阶因子分解机](https:\u002F\u002Fpapers.nips.cc\u002Fpaper\u002F6144-higher-order-factorization-machines.pdf)                                                                                                | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FHOFM)         | `torch`       |\n| 11  | IJCAI'17          | [DeepFM](.\u002Fmodel_zoo\u002FDeepFM)             | [DeepFM：基于因子分解机的神经网络用于点击率预测](https:\u002F\u002Farxiv.org\u002Fabs\u002F1703.04247) :triangular_flag_on_post:**华为**                                                                 | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDeepFM)       | `torch`, `tf` |\n| 12  | SIGIR'17          | [NFM](.\u002Fmodel_zoo\u002FNFM)                   | [稀疏预测分析中的神经因子分解机](https:\u002F\u002Fdl.acm.org\u002Fcitation.cfm?id=3080777)                                                                                                     | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FNFM)          | `torch`       |\n| 13  | IJCAI'17          | [AFM](.\u002Fmodel_zoo\u002FAFM)                   | [注意力因子分解机：通过注意力网络学习特征交互权重](http:\u002F\u002Fwww.ijcai.org\u002Fproceedings\u002F2017\u002F0435.pdf)                                                        | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FAFM)          | `torch`       |\n| 14  | ADKDD'17          | [DCN](.\u002Fmodel_zoo\u002FDCN)                   | [用于广告点击预测的深度与交叉网络](https:\u002F\u002Farxiv.org\u002Fabs\u002F1708.05123) :triangular_flag_on_post:**谷歌**                                                                                           | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDCN)          | `torch`, `tf` |\n| 15  | WWW'18            | [FwFM](.\u002Fmodel_zoo\u002FFwFM)                 | [展示广告中用于点击率预测的字段加权因子分解机](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1806.03514.pdf) :triangular_flag_on_post:**Oath, TouchPal, LinkedIn, 阿里巴巴**           | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFwFM)         | `torch`       |\n| 16  | KDD'18            | [xDeepFM](.\u002Fmodel_zoo\u002FxDeepFM)           | [xDeepFM：结合显式和隐式特征交互的推荐系统](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1803.05170.pdf) :triangular_flag_on_post:**微软**                                            | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FxDeepFM)      | `torch`       |\n| 17  | CIKM'19           | [FiGNN](.\u002Fmodel_zoo\u002FFiGNN)               | [FiGNN：通过图神经网络建模特征交互以进行点击率预测](https:\u002F\u002Farxiv.org\u002Fabs\u002F1910.05552)                                                                                           | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFiGNN)        | `torch`       |\n| 18  | CIKM'19           | [AutoInt\u002FAutoInt+](.\u002Fmodel_zoo\u002FAutoInt)  | [AutoInt：通过自注意力神经网络自动学习特征交互](https:\u002F\u002Farxiv.org\u002Fabs\u002F1810.11921)                                                                                          | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FAutoInt)      | `torch`       |\n| 19  | RecSys'19         | [FiBiNET](.\u002Fmodel_zoo\u002FFiBiNET)           | [FiBiNET：结合特征重要性和双线性特征交互的点击率预测](https:\u002F\u002Farxiv.org\u002Fabs\u002F1905.09433) :triangular_flag_on_post:**新浪微博**                            | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFiBiNET)      | `torch`       |\n| 20  | WWW'19            | [FGCNN](.\u002Fmodel_zoo\u002FFGCNN)               | [通过卷积神经网络生成特征以进行点击率预测](https:\u002F\u002Farxiv.org\u002Fabs\u002F1904.04447) :triangular_flag_on_post:**华为**                                                    | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFGCNN)        | `torch`       |\n| 21  | AAAI'19           | [HFM\u002FHFM+](.\u002Fmodel_zoo\u002FHFM)              | [全息因子分解机用于推荐](https:\u002F\u002Fojs.aaai.org\u002F\u002Findex.php\u002FAAAI\u002Farticle\u002Fview\u002F4448)                                                                                                 | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FHFM)          | `torch`       |\n| 22  | Arxiv'19          | [DLRM](.\u002Fmodel_zoo\u002FDLRM)                 | [用于个性化和推荐系统的深度学习推荐模型](https:\u002F\u002Farxiv.org\u002Fabs\u002F1906.00091) :triangular_flag_on_post:**Facebook**                                                     | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDLRM)         | `torch`       |\n| 23  | NeuralNetworks'20 | [ONN](.\u002Fmodel_zoo\u002FONN)                   | [操作感知神经网络用于用户响应预测](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1904.12579)                                                                                                                | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FONN)          | `torch`, `tf`      |\n| 24  | AAAI'20           | [AFN\u002FAFN+](.\u002Fmodel_zoo\u002FAFN)              | [自适应因子网络：学习自适应阶数的特征交互](https:\u002F\u002Fojs.aaai.org\u002Findex.php\u002FAAAI\u002Farticle\u002Fview\u002F5768)                                                                           | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FAFN)          | `torch`       |\n| 25  | AAAI'20           | [LorentzFM](.\u002Fmodel_zoo\u002FLorentzFM)       | [利用洛伦兹因子分解学习特征交互](https:\u002F\u002Farxiv.org\u002Fabs\u002F1911.09821) :triangular_flag_on_post:**eBay**                                                                               | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FLorentzFM)    | `torch`       |\n| 26  | WSDM'20           | [InterHAt](.\u002Fmodel_zoo\u002FInterHAt)         | [通过层次化注意力实现可解释的点击率预测](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002F10.1145\u002F3336191.3371785) :triangular_flag_on_post:**NEC Labs, 谷歌**                                      | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FInterHAt)     | `torch`       |\n| 27  | DLP-KDD'20        | [FLEN](.\u002Fmodel_zoo\u002FFLEN)                 | [FLEN：利用领域进行可扩展的点击率预测](https:\u002F\u002Farxiv.org\u002Fabs\u002F1911.04690) :triangular_flag_on_post:**腾讯**                                                                                     | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFLEN)         | `torch`       |\n| 28  | CIKM'20           | [DeepIM](.\u002Fmodel_zoo\u002FDeepIM)             | [深度交互机器：一种简单但有效的高阶特征交互模型](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002Fabs\u002F10.1145\u002F3340531.3412077) :triangular_flag_on_post:**阿里巴巴、RealAI**                   | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDeepIM)       | `torch`       |\n| 29  | WWW'21            | [FmFM](.\u002Fmodel_zoo\u002FFmFM)                 | [FM^2：用于推荐系统的字段矩阵化因子分解机](https:\u002F\u002Farxiv.org\u002Fabs\u002F2102.12994) :triangular_flag_on_post:**雅虎**                                                                      | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFmFM)         | `torch`       |\n| 30  | WWW'21            | [DCN-V2](.\u002Fmodel_zoo\u002FDCNv2)              | [DCN V2：改进的深度与交叉网络及面向Web规模排序学习的实用经验](https:\u002F\u002Farxiv.org\u002Fabs\u002F2008.13535) :triangular_flag_on_post:**谷歌**                                      | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDCNv2)        | `torch`       |\n| 31  | CIKM'21           | [DESTINE](.\u002Fmodel_zoo\u002FDESTINE)           | [解耦的自注意力神经网络用于点击率预测](https:\u002F\u002Farxiv.org\u002Fabs\u002F2101.03654) :triangular_flag_on_post:**阿里巴巴**                                                          | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDESTINE)      | `torch`       |\n| 32  | CIKM'21           | [EDCN](.\u002Fmodel_zoo\u002FEDCN)                 | [通过信息共享增强显式和隐式特征交互，用于并行深度点击率模型](https:\u002F\u002Fdlp-kdd.github.io\u002Fassets\u002Fpdf\u002FDLP-KDD_2021_paper_12.pdf) :triangular_flag_on_post:**华为** | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FEDCN)         | `torch`       |\n| 33  | DLP-KDD'21        | [MaskNet](.\u002Fmodel_zoo\u002FMaskNet)           | [MaskNet：通过实例引导的掩码将逐特征乘法引入点击率排序模型](https:\u002F\u002Farxiv.org\u002Fabs\u002F2102.07619) :triangular_flag_on_post:**新浪微博**                                      | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FMaskNet)      | `torch`       |\n| 34  | SIGIR'21          | [SAM](.\u002Fmodel_zoo\u002FSAM)                   | [再次审视点击率预测：注意力就是你所需要的吗？](https:\u002F\u002Farxiv.org\u002Fabs\u002F2105.05563) :triangular_flag_on_post:**BOSS Zhipin**                                                                        | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FSAM)          | `torch`       |\n| 35  | KDD'21            | [AOANet](.\u002Fmodel_zoo\u002FAOANet)             | [用于在线推荐的架构与操作自适应网络](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002F10.1145\u002F3447548.3467133) :triangular_flag_on_post:**滴滴出行**                                              | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FAOANet)       | `torch`       |\n| 36  | AAAI'23           | [FinalMLP](.\u002Fmodel_zoo\u002FFinalMLP)         | [FinalMLP：用于点击率预测的增强型双流MLP模型](https:\u002F\u002Farxiv.org\u002Fabs\u002F2304.00902) :triangular_flag_on_post:**华为**                                                                                                               |     [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFinalMLP)         | `torch`       |\n| 37  | SIGIR'23          | [FinalNet](.\u002Fmodel_zoo\u002FFinalNet)               | [FINAL：用于点击率预测的因子化交互层](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002F10.1145\u002F3539618.3591988) :triangular_flag_on_post:**华为**                                                                                                               |     [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FFinalNet)         | `torch`       |\n| 38  | SIGIR'23          | [EulerNet](.\u002Fmodel_zoo\u002FEulerNet)               | [EulerNet：通过欧拉公式自适应地学习特征交互以进行点击率预测](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002F10.1145\u002F3539618.3591681) :triangular_flag_on_post:**华为**                                                                                                               |     [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002FEthan-TZ\u002FEulerNet\u002Ftree\u002Fmain\u002F%23Code4FuxiCTR%23)         | `torch`       |\n| 39  | CIKM'23           | [GDCN](.\u002Fmodel_zoo\u002FGDCN)         | [迈向更深、更轻量且可解释的交叉网络以进行点击率预测](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002Fpdf\u002F10.1145\u002F3583780.3615089) :triangular_flag_on_post:**微软**                                                                                                               |           | `torch`       |\n| 40  | ICML'24          | [WuKong](.\u002Fmodel_zoo\u002FWuKong)               | [悟空：迈向大规模推荐的规模法则](https:\u002F\u002Farxiv.org\u002Fabs\u002F2403.02545) :triangular_flag_on_post:**Meta**                                                        |   [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FWuKong)    | `torch`       |\n|\u003Ctr>\u003Cth colspan=6 align=\"center\">:open_file_folder: **行为序列建模**\u003C\u002Fth>\u003C\u002Ftr>|\n| 42  | KDD'18            | [DIN](.\u002Fmodel_zoo\u002FDIN)                   | [用于点击率预测的深度兴趣网络](https:\u002F\u002Fwww.kdd.org\u002Fkdd2018\u002Faccepted-papers\u002Fview\u002Fdeep-interest-network-for-click-through-rate-prediction) :triangular_flag_on_post:**阿里巴巴**        |   [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDIN)       | `torch`       |\n| 43  | AAAI'19           | [DIEN](.\u002Fmodel_zoo\u002FDIEN)                 | [用于点击率预测的深度兴趣演化网络](https:\u002F\u002Farxiv.org\u002Fabs\u002F1809.03672) :triangular_flag_on_post:**阿里巴巴**                                                                      |   [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDIEN)        | `torch`       |\n| 44  | DLP-KDD'19        | [BST](.\u002Fmodel_zoo\u002FBST)                   | [阿里巴巴电商推荐中的行为序列Transformer](https:\u002F\u002Farxiv.org\u002Fabs\u002F1905.06874) :triangular_flag_on_post:**阿里巴巴**                                                                 |  [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FBST)     | `torch`       |\n| 45  | CIKM'20           | [DMIN](.\u002Fmodel_zoo\u002FDMIN)                 | [用于点击率预测的深度多兴趣网络](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002F10.1145\u002F3340531.3412092) :triangular_flag_on_post:**阿里巴巴**                                                            | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDMIN)                                                                                                                 | `torch`       |\n| 46  | AAAI'20           | [DMR](.\u002Fmodel_zoo\u002FDMR)                   | [用于个性化点击率预测的深度匹配排序模型](https:\u002F\u002Fojs.aaai.org\u002Findex.php\u002FAAAI\u002Farticle\u002Fview\u002F5346) :triangular_flag_on_post:**阿里巴巴**                                           |    [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDMR)                                                                                                                  | `torch`       |\n| 47  | KDD'23           | [TransAct](.\u002Fmodel_zoo\u002FTransAct)                 | [TransAct：Pinterest 推荐中的基于 Transformer 的实时用户行为模型](https:\u002F\u002Farxiv.org\u002Fabs\u002F2306.00248) :triangular_flag_on_post:**Pinterest**                                                       | [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FTransAct)         | `torch`       |\n|\u003Ctr>\u003Cth colspan=6 align=\"center\">:open_file_folder: **长序列建模**\u003C\u002Fth>\u003C\u002Ftr>|\n| 48  | CIKM'20          | [SIM](.\u002Fmodel_zoo\u002FLongCTR\u002FSIM)                   | [基于搜索的用户兴趣建模，结合终身序列行为数据进行点击率预测](https:\u002F\u002Farxiv.org\u002Fabs\u002F2006.05639) :triangular_flag_on_post:**阿里巴巴**                                                               |                                                                                                                 | `torch`       |\n| 49  | DLP-KDD'22          | [ETA](.\u002Fmodel_zoo\u002FLongCTR\u002FETA)                   | [高效长序列用户数据建模以进行点击率预测](https:\u002F\u002Farxiv.org\u002Fabs\u002F2209.12212) :triangular_flag_on_post:**阿里巴巴**                                                               |                                                                                                                 | `torch`       |\n| 50  | CIKM'22           | [SDIM](.\u002Fmodel_zoo\u002FLongCTR\u002FSDIM)                 | [采样就是你在点击率预测中对长期用户行为建模所需要的全部](https:\u002F\u002Farxiv.org\u002Fabs\u002F2205.10249) :triangular_flag_on_post:**美团**                                                       |                                                                                                                 | `torch`       |\n| 51  | KDD'23           | [TWIN](.\u002Fmodel_zoo\u002FLongCTR\u002FTWIN)                 | [TWIN：快手点击率预测中用于终身用户行为建模的两阶段兴趣网络](https:\u002F\u002Farxiv.org\u002Fabs\u002F2302.02352) :triangular_flag_on_post:**快手**                                                       |                                                                                                                 | `torch`       |\n| 52  | KDD'25           | [MIRRN](.\u002Fmodel_zoo\u002FLongCTR\u002FMIRRN)                 | [多粒度兴趣检索与精炼网络用于点击率预测中的长期用户行为建模](https:\u002F\u002Farxiv.org\u002Fabs\u002F2411.15005) :triangular_flag_on_post:**华为**                                                       |                                                                                                                 | `torch`       |\n|\u003Ctr>\u003Cth colspan=6 align=\"center\">:open_file_folder: **动态权重网络**\u003C\u002Fth>\u003C\u002Ftr>|\n| 53  | NeurIPS'22          | [APG](.\u002Fmodel_zoo\u002FAPG)               | [APG：用于点击率预测的自适应参数生成网络](https:\u002F\u002Farxiv.org\u002Fabs\u002F2203.16218) :triangular_flag_on_post:**阿里巴巴**                                |    [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FAPG)                                                                                                   | `torch`       |\n| 54  | KDD'23        | [PPNet](.\u002Fmodel_zoo\u002FPEPNet)             | [PEPNet：用于注入个性化先验信息的参数和嵌入个性化网络](https:\u002F\u002Farxiv.org\u002Fabs\u002F2302.01115) :triangular_flag_on_post:**快手**                          |    [:arrow_upper_right:](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FPPNet)                                                                                                   | `torch`       |\n|\u003Ctr>\u003Cth colspan=6 align=\"center\">:open_file_folder: **多任务建模**\u003C\u002Fth>\u003C\u002Ftr>|\n| 55  |     Arxiv'17      | [ShareBottom](.\u002Fmodel_zoo\u002Fmultitask\u002FShareBottom)               | [深度神经网络中多任务学习概述](https:\u002F\u002Farxiv.org\u002Fabs\u002F1706.05098)                                                                                            |                                                                                                                 | `torch`       |\n| 56  | KDD'18          | [MMoE](.\u002Fmodel_zoo\u002Fmultitask\u002FMMOE)               | [利用多门专家混合体在多任务学习中建模任务关系](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002Fpdf\u002F10.1145\u002F3219819.3220007) :triangular_flag_on_post:**谷歌**                                                                                            |                                                                                                                 | `torch`       |\n| 57  | RecSys'20          | [PLE](.\u002Fmodel_zoo\u002Fmultitask\u002FPLE)               | [渐进式分层提取 (PLE)：一种用于个性化推荐的新颖多任务学习 (MTL) 模型](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002F10.1145\u002F3383313.3412236) :triangular_flag_on_post:**腾讯**                                                                                            |                                                                                                                 | `torch`       |\n\n## 基准测试\n\n我们已在一组开放数据集上对 FuxiCTR 模型进行了基准测试，具体如下：\n\n+ :star: [CTR 预测的基准数据集](https:\u002F\u002Fgithub.com\u002Freczoo\u002FDatasets?tab=readme-ov-file#ctr-prediction)\n+ :star: [基准设置与运行步骤](https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr)\n+ :star: [CTR 预测的基准排行榜](https:\u002F\u002Fopenbenchmark.github.io\u002FBARS\u002FCTR\u002Fleaderboard)\n\n## 依赖项\n\nFuxiCTR 具有以下依赖项：\n\n+ Python 3.9+\n+ PyTorch 1.10.0–2.1.2（若用于 PyTorch 模型）\n+ TensorFlow 2.1（若用于 TensorFlow 模型）\n\n请通过 `pip install -r requirements.txt` 安装其他所需软件包。\n\n## 快速入门\n\n1. 运行示例代码\n\n   demo 目录中提供了示例代码，展示了 FuxiCTR 的一些基本用法。用户可以运行这些示例以快速上手并了解工作流程。\n   \n   ```\n   cd demo\n   python example1_build_dataset_to_parquet.py\n   python example2_DeepFM_with_parquet_input.py\n   ```\n\n2. 在小型数据集上运行模型\n\n   用户可以按照以下命令轻松运行模型库中的每个模型，这里以运行 DCN 为例。此外，用户还可以修改数据集配置文件和模型配置文件，以在自己的数据集上运行或使用新的超参数。更多详细信息请参阅 [README](.\u002Fmodel_zoo\u002FDCN\u002FDCN_torch\u002FREADME.md)。\n   \n   ```\n   cd model_zoo\u002FDCN\u002FDCN_torch\n   python run_expid.py --expid DCN_test --gpu 0\n\n   # 根据目标模型名称更改 `MODEL`\n   cd model_zoo\u002FMODEL\n   python run_expid.py --expid MODEL_test --gpu 0\n   ```\n\n3. 在基准数据集上运行模型（例如 Criteo）\n\n   用户可以参考[基准测试部分](#Benchmarking)，获取基准数据集和运行步骤，以复现现有结果。请参阅此处示例：https:\u002F\u002Fgithub.com\u002Freczoo\u002FBARS\u002Ftree\u002Fmain\u002Franking\u002Fctr\u002FDCNv2\u002FDCNv2_criteo_x1\n\n4. 实现新模型\n\n   FuxiCTR 库设计为模块化结构，因此用户可以根据自身需求覆盖各个组件。在许多情况下，只需实现模型类即可创建自定义模型。如果数据预处理或数据加载器不适用，也可以通过[核心 API](https:\u002F\u002Fwww.processon.com\u002Fview\u002Flink\u002F63cfcfab4e30670eac4a81c7) 自行实现新的版本。我们提供了一个具体示例，实现了我们最近在 AAAI 2023 上发表的新模型 [FinalMLP](https:\u002F\u002Freczoo.github.io\u002FFinalMLP)。\n\n5. 调整模型超参数\n\n   FuxiCTR 目前支持使用多 GPU 对模型超参数进行快速网格搜索。以下示例展示了使用 4 个 GPU 进行 8 组实验的网格搜索。\n   \n   ```\n   cd experiment\n   python run_param_tuner.py --config config\u002FDCN_tiny_parquet_tuner_config.yaml --gpu 0 1 2 3 0 1 2 3\n   ```\n\n## 🔥 引用\n\n如果您在公开研究中使用了我们的代码或基准测试，请引用以下两篇论文。\n\n+ Jieming Zhu, Quanyu Dai, Liangcai Su, Rong Ma, Jinyang Liu, Guohao Cai, Xi Xiao, Rui Zhang. [BARS：迈向推荐系统的开放基准测试](https:\u002F\u002Farxiv.org\u002Fabs\u002F2205.09626)。*第 45 届国际 ACM SIGIR 信息检索研究与发展会议 (SIGIR)*，2022 年。[[Bibtex](https:\u002F\u002Fdblp.org\u002Frec\u002Fconf\u002Fsigir\u002FZhuDSMLCXZ22.html?view=bibtex)]\n+ Jieming Zhu, Jinyang Liu, Shuai Yang, Qi Zhang, Xiuqiang He. [BARS-CTR：点击率预测的开放基准测试](https:\u002F\u002Farxiv.org\u002Fabs\u002F2009.05794)。*第 30 届 ACM 国际信息与知识管理会议 (CIKM)*，2021 年。[[Bibtex](https:\u002F\u002Fdblp.org\u002Frec\u002Fconf\u002Fcikm\u002FZhuLYZH21.html?view=bibtex)]\n\n## 🙋 讨论\n\n欢迎加入我们的微信群，提出任何问题或参与讨论。如果您对推荐系统的研究与实践感兴趣，请通过微信群联系我们。\n\n![扫描二维码](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Freczoo_FuxiCTR_readme_b19ca98f40a7.jpg)","# FuxiCTR 快速上手指南\n\nFuxiCTR 是一个专注于点击率（CTR）预测的开源工具库，支持高度可配置、自动调参和实验复现。它同时支持 PyTorch 和 TensorFlow 框架，内置了从经典 LR、FM 到前沿 DeepFM、DCN、AutoInt 等数十种主流模型。\n\n## 1. 环境准备\n\n在开始之前，请确保您的开发环境满足以下要求：\n\n*   **操作系统**：Linux \u002F macOS \u002F Windows\n*   **Python 版本**：3.9 或更高版本\n*   **深度学习框架**（任选其一或同时安装）：\n    *   PyTorch 1.10+\n    *   TensorFlow 2.1+\n*   **其他依赖**：`numpy`, `pandas`, `scikit-learn` 等（安装 FuxiCTR 时会自动处理大部分依赖）\n\n> **国内加速建议**：\n> 推荐使用国内镜像源安装 Python 依赖，以提升下载速度。\n> *   PyTorch 国内镜像：https:\u002F\u002Fmirror.tuna.tsinghua.edu.cn\u002Fhelp\u002Fpytorch\u002F\n> *   pip 临时使用清华源：`pip install -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple \u003Cpackage_name>`\n\n## 2. 安装步骤\n\n您可以通过 PyPI 直接安装稳定版，或从源码安装最新版。\n\n### 方式一：通过 PyPI 安装（推荐）\n\n```bash\npip install fuxictr\n```\n\n若需指定框架版本或使用国内镜像加速：\n\n```bash\npip install -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple fuxictr\n```\n\n### 方式二：从源码安装\n\n如果您需要最新的功能或贡献代码，可以克隆仓库进行安装：\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Freczoo\u002FFuxiCTR.git\ncd FuxiCTR\npip install -e .\n```\n\n## 3. 基本使用\n\nFuxiCTR 的核心设计理念是“配置驱动”。用户只需编写配置文件（YAML 格式），即可轻松运行数据预处理、模型训练和评估流程，无需编写大量重复代码。\n\n### 第一步：准备数据集\n\nFuxiCTR 支持多种公开数据集（如 Criteo, Avazu, MovieLens）。您需要将数据转换为 FuxiCTR 支持的格式，或者使用内置的数据加载器。通常数据文件包含特征列和标签列。\n\n### 第二步：编写配置文件\n\n创建一个名为 `model_config.yaml` 的文件。以下是一个使用 **DeepFM** 模型在 **Criteo** 数据集上运行的最小化配置示例：\n\n```yaml\n# model_config.yaml\ndataset_id: criteo_sample\ndata_root: ..\u002Fdata\u002F\nfeature_config: ..\u002Fdata\u002Fcriteo_sample_features.json\n\nmodel_root: .\u002Fcheckpoints\u002F\ngpu_ids: [0]\n\nmodel: DeepFM\ntask: binary_classification\nmetrics: ['logloss', 'auc']\n\nnet_config:\n    embedding_dim: 10\n    hidden_units: [64, 32]\n    dropout_rates: [0.1, 0.1]\n    batch_norm: True\n\ntrain_config:\n    epoch: 10\n    batch_size: 4096\n    learning_rate: 1.0e-3\n    optimizer: adam\n    loss: binary_crossentropy\n```\n\n### 第三步：运行实验\n\n使用 `fuxictr` 命令行工具加载配置并启动训练。以下命令将执行数据预处理（如果尚未完成）并训练模型：\n\n```bash\npython run_experiments.py --config model_config.yaml --mode train\n```\n\n训练完成后，您可以在 `.\u002Fcheckpoints\u002F` 目录下找到保存的模型权重，并在终端查看评估指标（如 AUC 和 LogLoss）。\n\n### 进阶：快速切换模型\n\n得益于其模块化设计，您只需修改配置文件中的 `model` 字段，即可轻松切换至其他模型（如 `DCN`, `AutoInt`, `xDeepFM` 等），而无需更改任何代码逻辑：\n\n```yaml\n# 将模型更改为 DCN\nmodel: DCN\nnet_config:\n    embedding_dim: 10\n    cross_layers: 3\n    # ... 其他参数\n```\n\n再次运行 `run_experiments.py` 即可复现新模型的基准结果。","某电商平台的推荐算法团队正面临紧急任务，需要在两周内为“双 11\"大促上线一套新的点击率（CTR）预估模型，以提升广告转化率。\n\n### 没有 FuxiCTR 时\n- **重复造轮子效率低**：团队成员需手动重写 LR、FM、DeepFM 等经典模型的底层代码，数据预处理逻辑分散且难以复用，耗费大量时间在基础构建上。\n- **调参复现困难**：不同成员使用的超参数配置格式不统一，导致实验结果无法横向对比，甚至出现同一模型在不同环境下表现不一致的“黑盒”问题。\n- **框架迁移成本高**：团队内部部分成员熟悉 PyTorch，部分习惯 TensorFlow，代码无法互通，维护两套技术栈增加了协作摩擦和出错风险。\n- **新模型验证慢**：想要尝试最新的特征交互模型时，缺乏标准化的基准测试流程，从阅读论文到代码落地往往需要数周时间。\n\n### 使用 FuxiCTR 后\n- **模块化快速搭建**：利用 FuxiCTR 预置的模型库（Model Zoo），通过配置文件即可一键调用 LR、DSSM 等数十种成熟模型，将模型开发周期从周缩短至小时级。\n- **实验可复现可控**：借助其统一的配置化调参机制，所有实验记录清晰可查，确保了基准测试的严格复现，让团队能精准定位性能提升的来源。\n- **双框架无缝支持**：FuxiCTR 同时支持 PyTorch 和 TensorFlow，团队成员可根据偏好选择框架，底层逻辑保持一致，极大降低了协作门槛。\n- **灵活扩展新算法**：基于其高扩展性架构，团队能快速将最新论文中的创新模型集成到现有流水线中，并在标准数据集上立即验证效果。\n\nFuxiCTR 通过标准化、配置化和可复现的特性，将算法团队从繁琐的工程实现中解放出来，使其能专注于核心策略优化与业务价值提升。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Freczoo_FuxiCTR_2478d4f9.png","reczoo","RECZOO","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Freczoo_dd8f065c.png","",null,"https:\u002F\u002Fgithub.com\u002Freczoo",[82,86],{"name":83,"color":84,"percentage":85},"Python","#3572A5",99.3,{"name":87,"color":88,"percentage":89},"Shell","#89e051",0.7,1395,224,"2026-04-05T02:02:29","Apache-2.0","未说明（依赖 PyTorch\u002FTensorFlow 的默认 GPU 支持）","未说明",{"notes":97,"python":98,"dependencies":99},"该工具同时支持 PyTorch 和 TensorFlow 框架，用户可根据需要选择安装其中一个或两个都安装。具体模型实现可能仅依赖其中一种框架（如表格中标注的 `torch` 或 `tf`）。README 中未明确提及操作系统、GPU 型号及内存的具体需求，通常取决于所选深度学习框架的要求及运行模型的数据规模。","3.9+",[100,101],"torch>=1.10","tensorflow>=2.1",[14,13],[104,105,106,107,108],"ctr-prediction","recommender-systems","ctr","cvr","pytorch","2026-03-27T02:49:30.150509","2026-04-06T08:45:21.292059",[112,117,122,127,131,136],{"id":113,"question_zh":114,"answer_zh":115,"source_url":116},17503,"运行大型数据集（如 taobaoad）时出现内存溢出（Memory Error）怎么办？","可以通过在配置文件中设置 \"data_block_size\" 参数将数据集分割成更小的部分来解决。例如，设置 \"data_block_size: 100000\"。这可以避免一次性加载过多数据导致内存不足。具体配置示例可参考相关 YAML 文件中的设置。","https:\u002F\u002Fgithub.com\u002Freczoo\u002FFuxiCTR\u002Fissues\u002F124",{"id":118,"question_zh":119,"answer_zh":120,"source_url":121},17504,"复现 TransAct 模型时，验证阶段的 Embedding 权重全变为 NaN 是什么原因？","这是因为演示数据中缺乏行为数据，导致整个序列被掩码（masked）。解决方法是修改代码中的掩码调整逻辑，确保并非所有 token 都被掩码。请在 `model_zoo\u002FTransAct\u002Fsrc\u002FTransAct.py` 第 223 行附近，确保调用 `self.adjust_mask(mask)` 来正确处理掩码，防止全序列被屏蔽。","https:\u002F\u002Fgithub.com\u002Freczoo\u002FFuxiCTR\u002Fissues\u002F101",{"id":123,"question_zh":124,"answer_zh":125,"source_url":126},17505,"如何将框架从二分类任务修改为多分类任务并添加相应的评价指标？","需要修改三个部分：1. 损失函数（loss）；2. 标签获取函数（get_label）；3. 评价指标（metrics）。通过调整这三处代码，即可支持多分类任务并使用对应的评估指标。","https:\u002F\u002Fgithub.com\u002Freczoo\u002FFuxiCTR\u002Fissues\u002F91",{"id":128,"question_zh":129,"answer_zh":130,"source_url":126},17506,"运行数据预处理脚本后生成的文件没有 .npz 后缀，或者加上后缀后报错 \"allow_pickle=False\" 怎么办？","这是版本兼容性问题。建议暂时使用 v2.2.3 版本来处理 npz 格式的数据，因为当前最新版本可能尚未完全适配该格式的读写操作。",{"id":132,"question_zh":133,"answer_zh":134,"source_url":135},17507,"文档中提到的 `na_value` 参数在代码中似乎未生效，应该使用哪个参数处理缺失值？","代码实际使用的是 `fill_na` 参数来处理缺失值，而非文档中提到的 `na_value`。这是一种文档与代码不一致的情况。建议在配置文件中使用 `fill_na` 来指定缺失值的填充策略，以确保特征预处理和分词过程正常工作。","https:\u002F\u002Fgithub.com\u002Freczoo\u002FFuxiCTR\u002Fissues\u002F144",{"id":137,"question_zh":138,"answer_zh":139,"source_url":140},17508,"在配置自定义特征处理器（user_defined_processor）时，是否需要为每一列单独指定列名参数？","是的，当前版本的设计要求为每一列单独指定配置。虽然这可能导致配置文件略显冗余，但这是为了确保处理器能正确作用于特定列。直接在配置中硬编码列名可能会导致如 `copy_from` 等功能出错，因此推荐按列分别配置。","https:\u002F\u002Fgithub.com\u002Freczoo\u002FFuxiCTR\u002Fissues\u002F105",[]]