[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-facebookresearch--Replica-Dataset":3,"tool-facebookresearch--Replica-Dataset":64},[4,17,27,35,48,56],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":16},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},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",[14,43,44,45,15,46,26,13,47],"数据工具","视频","插件","其他","音频",{"id":49,"name":50,"github_repo":51,"description_zh":52,"stars":53,"difficulty_score":10,"last_commit_at":54,"category_tags":55,"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",[15,14,13,26,46],{"id":57,"name":58,"github_repo":59,"description_zh":60,"stars":61,"difficulty_score":10,"last_commit_at":62,"category_tags":63,"status":16},519,"PaddleOCR","PaddlePaddle\u002FPaddleOCR","PaddleOCR 是一款基于百度飞桨框架开发的高性能开源光学字符识别工具包。它的核心能力是将图片、PDF 等文档中的文字提取出来，转换成计算机可读取的结构化数据，让机器真正“看懂”图文内容。\n\n面对海量纸质或电子文档，PaddleOCR 解决了人工录入效率低、数字化成本高的问题。尤其在人工智能领域，它扮演着连接图像与大型语言模型（LLM）的桥梁角色，能将视觉信息直接转化为文本输入，助力智能问答、文档分析等应用场景落地。\n\nPaddleOCR 适合开发者、算法研究人员以及有文档自动化需求的普通用户。其技术优势十分明显：不仅支持全球 100 多种语言的识别，还能在 Windows、Linux、macOS 等多个系统上运行，并灵活适配 CPU、GPU、NPU 等各类硬件。作为一个轻量级且社区活跃的开源项目，PaddleOCR 既能满足快速集成的需求，也能支撑前沿的视觉语言研究，是处理文字识别任务的理想选择。",74939,"2026-04-05T23:16:38",[26,14,13,46],{"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":80,"owner_url":81,"languages":82,"stars":102,"forks":103,"last_commit_at":104,"license":105,"difficulty_score":10,"env_os":106,"env_gpu":107,"env_ram":107,"env_deps":108,"category_tags":117,"github_topics":79,"view_count":10,"oss_zip_url":79,"oss_zip_packed_at":79,"status":16,"created_at":118,"updated_at":119,"faqs":120,"releases":151},611,"facebookresearch\u002FReplica-Dataset","Replica-Dataset","The Replica Dataset v1 as published in https:\u002F\u002Farxiv.org\u002Fabs\u002F1906.05797 .","Replica-Dataset 是一个专注于室内空间的高精度三维重建数据集。它收录了多个真实室内场景的数字副本，提供清洁的稠密几何结构、高动态范围纹理，以及玻璃和镜面表面的详细参数。此外，数据集中还包含了平面分割、语义类别及实例分割信息，极大地丰富了场景的理解维度。\n\n这一数据集主要解决了人工智能与机器人领域在虚拟环境中缺乏逼真室内数据的问题。通过提供标准化的格式，它让研究者无需从零开始采集和标注即可进行实验。Replica-Dataset 非常适合计算机视觉研究人员、AI 开发者以及从事仿真模拟的工程师。其独特之处在于内置了兼容 AI Habitat 框架的导出功能，能够无缝支持智能体训练任务。同时，配套的 SDK 允许用户进行可视化检查和无头渲染，为场景分析与算法验证提供了便利。无论是探索 SLAM 技术还是训练导航智能体，这里都是理想的起点。","# Replica Dataset\n\nThe Replica Dataset is a dataset of high quality reconstructions of a\nvariety of indoor spaces. Each reconstruction has clean dense geometry, high\nresolution and high dynamic range textures, glass and mirror surface\ninformation, planar segmentation as well as\nsemantic class and instance segmentation.\nSee the [technical report](https:\u002F\u002Farxiv.org\u002Fabs\u002F1906.05797) for more details.\n\n![Replica Modalities](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Ffacebookresearch_Replica-Dataset_readme_11418741ea3b.png)\n\nThe Replica SDK contained in this repository allows visual inspection of the\ndatasets via the ReplicaViewer and gives an example of how to render out images\nfrom the scenes headlessly via the ReplicaRenderer. \n\nFor machine learning purposes each dataset also contains an export to the format\nemployed by [AI Habitat](https:\u002F\u002Fwww.aihabitat.org\u002F) and is therefore usable\nseamlessly in that framework for AI agent training and other ML tasks.\n\n## Citing the Replica Dataset\n\nIf you use the Replica dataset in your research directly or indirectly via derivative datasets or frameworks, please cite the following\n[technical report](https:\u002F\u002Farxiv.org\u002Fabs\u002F1906.05797):\n```\n@article{replica19arxiv,\n  title =   {The {R}eplica Dataset: A Digital Replica of Indoor Spaces},\n  author =  {Julian Straub and Thomas Whelan and Lingni Ma and Yufan Chen and Erik Wijmans and Simon Green and Jakob J. Engel and Raul Mur-Artal and Carl Ren and Shobhit Verma and Anton Clarkson and Mingfei Yan and Brian Budge and Yajie Yan and Xiaqing Pan and June Yon and Yuyang Zou and Kimberly Leon and Nigel Carter and Jesus Briales and  Tyler Gillingham and  Elias Mueggler and Luis Pesqueira and Manolis Savva and Dhruv Batra and Hauke M. Strasdat and Renzo De Nardi and Michael Goesele and Steven Lovegrove and Richard Newcombe },\n  journal = {arXiv preprint arXiv:1906.05797},\n  year =    {2019}\n}\n```\n\n## Replica Dataset\n\nThe following 18 scenes are included in this initial release:\n\n![Replica Dataset](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Ffacebookresearch_Replica-Dataset_readme_c3098187eb10.png)\n\n![Replica Dataset](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Ffacebookresearch_Replica-Dataset_readme_6a9a398b139f.png)\n\nEach Replica contains the following assets:\n```\n├── glass.sur\n├── habitat\n    ├── mesh_semantic.ply \n    ├── mesh_semantic.navmesh\n    ├── info_semantic.json\n    ├── mesh_preseg_semantic.ply \n    ├── mesh_preseg_semantic.navmesh\n    ├── info_preseg_semantic.json\n    ├── replica_stage.stage_config.json\n    └── sorted_faces.bin \n├── mesh.ply\n├── preseg.bin\n├── preseg.json\n├── semantic.bin\n├── semantic.json\n└── textures\n    ├── 0-color-ptex.hdr\n    ├── 0-color-ptex.w\n    ├── 1-color-ptex.hdr\n    ├── 1-color-ptex.w\n    ├── ...\n    └── parameters.json\n```\nThe different files contain the following:\n- `glass.sur`: parameterization of glass and mirror surfaces.\n- `mesh.ply`: the quad mesh of the scene with vertex colors.\n- `preseg.json` and `preseg.bin`: the presegmentation in terms of planes and non-planes of the scene.\n- `semantic.json` and `semantic.bin`: the semantic segmentation of the scene.\n- `textures`: the high resolution and high dynamic range textures of the scene.\n- `habitat\u002Fmesh*semantic.ply`: the quad meshes including semantic or presegmentation information for AI Habitat. \n- `habitat\u002Finfo*semantic.json`: mapping from instance IDs in the respective `mesh_*.ply` to semantic names.\n- `habitat\u002Fmesh*semantic.navmesh`: navigation grid for AI Habitat.\n- `habitat\u002Freplica_stage.stage_config.json`: configuration file defining scene level parameters for habitat-sim.\n- `habitat\u002Fsorted_faces.bin`: binary file containing pre-processed geometry data for habitat-sim Ptex rendering support.\n\n### Download on Mac OS and Linux\nMake sure `pigz`, `wget`, and `unzip` are installed:\n```\n# on Mac OS\nbrew install wget pigz unzip\n# on Ubuntu\nsudo apt-get install wget pigz unzip\n```\nTo download and decompress the dataset use the `download.sh` script:\n```\n.\u002Fdownload.sh \u002Fpath\u002Fto\u002Freplica_v1\n```\n\n### Download on Windows\n\nExecute `win_download.bat` to download Replica.\n\n## Replica SDK\n\n### Setup\nAfter installing the dependencies of [Pangolin](https:\u002F\u002Fgithub.com\u002Fstevenlovegrove\u002FPangolin),\nthe Replica SDK can be compiled using the build script via\n```\ngit submodule update --init\n.\u002Fbuild.sh\n```\nIt requires the dependencies of\n[Pangolin](https:\u002F\u002Fgithub.com\u002Fstevenlovegrove\u002FPangolin) and\n[Eigen](https:\u002F\u002Fgithub.com\u002Feigenteam\u002Feigen-git-mirror)\nto be installed. If you wish to use the headless renderer ensure you have the libegl1-mesa-dev package.\n\n### ReplicaViewer\n\nReplicaViewer is an interactive UI to explore the Replica Dataset. \n\n```\n.\u002Fbuild\u002Fbin\u002FReplicaViewer mesh.ply \u002Fpath\u002Fto\u002Fatlases [mirrorFile]\n```\n\n![ReplicaViewer](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Ffacebookresearch_Replica-Dataset_readme_49dd4ff82dd5.png)\n\nThe exposure value for rendering from the HDR textures can be adjusted on the\ntop left. \n\n### ReplicaRenderer\n\nThe ReplicaRenderer shows how to render out images from a Replica for a\nprogrammatically defined trajectory without UI. This executable can be run\nheadless on a server if so desired. \n\n```\n.\u002Fbuild\u002Fbin\u002FReplicaRenderer mesh.ply textures glass.sur\n```\n\n## Replica and AI Habitat\n\nTo use Replica within AI Habitat checkout the AI Habitat Sim at [https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fhabitat-sim](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fhabitat-sim).\nAfter building the project you can launch the test viewer to verify that everything works:\n```\n.\u002Fbuild\u002Fviewer --dataset \u002FPATH\u002FTO\u002FREPLICA\u002Freplica.scene_dataset_config.json -- frl_apartment_0\n```\n\n## Team \n\nJulian Straub,  Thomas Whelan, Lingni Ma, Yufan Chen, Erik Wijmans, Simon Green, Jakob J. Engel, Raul Mur-Artal, Carl Ren, Shobhit Verma, Anton Clarkson, Mingfei Yan, Brian Budge, Yajie Yan, Xiaqing Pan, June Yon, Yuyang Zou, Kimberly Leon, Nigel Carter, Jesus Briales,  Tyler Gillingham Elias Mueggler, Luis Pesqueira, Manolis Savva, Dhruv Batra, Hauke M. Strasdat, Renzo De Nardi, Michael Goesele, Steven Lovegrove, and Richard Newcombe.\n\n## Contact\n\n[Julian.Straub@oculus.com](Julian.Straub@oculus.com)\n\n## Acknowledgements\n\nThe Replica dataset would not have been possible without the hard work and contributions of Matthew Banks, Christopher Dotson, Rashad Barber, Justin Blosch, Ethan Henderson, Kelley Greene, Michael Thot, Matthew Winterscheid, Robert Johnston, Abhijit Kulkarni, Robert Meeker, Jamie Palacios, Tony Phan, Tim Petrvalsky, Sayed Farhad Sadat, Manuel Santana, Suruj Singh, Swati Agrawal, and Hannah Woolums.\n\n## License\n\nSee the LICENSE file for details.\n","# Replica 数据集\n\nReplica 数据集 (Replica Dataset) 是一个包含多种室内空间高质量重建的数据集。每个重建都拥有干净的稠密几何结构、高分辨率和高动态范围 (HDR) 纹理、玻璃和镜面表面信息、平面分割 (Planar Segmentation) 以及语义类别和实例分割 (Semantic Class and Instance Segmentation)。更多详情请参见 [技术报告](https:\u002F\u002Farxiv.org\u002Fabs\u002F1906.05797)。\n\n![Replica Modalities](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Ffacebookresearch_Replica-Dataset_readme_11418741ea3b.png)\n\n本仓库中包含的 Replica SDK（软件开发工具包）允许通过 ReplicaViewer 对数据集进行视觉检查，并通过 ReplicaRenderer 提供了如何无头模式 (Headless) 渲染场景图像的示例。 \n\n出于机器学习 (Machine Learning) 目的，每个数据集还包含导出到 [AI Habitat](https:\u002F\u002Fwww.aihabitat.org\u002F) 所使用的格式，因此可以无缝地在该框架中用于 AI 代理 (AI Agent) 训练和其他机器学习 (ML) 任务。\n\n## 引用 Replica 数据集\n\n如果您在研究中直接使用或间接通过衍生数据集或框架使用 Replica 数据集，请引用以下 [技术报告](https:\u002F\u002Farxiv.org\u002Fabs\u002F1906.05797)：\n```\n@article{replica19arxiv,\n  title =   {The {R}eplica Dataset: A Digital Replica of Indoor Spaces},\n  author =  {Julian Straub and Thomas Whelan and Lingni Ma and Yufan Chen and Erik Wijmans and Simon Green and Jakob J. Engel and Raul Mur-Artal and Carl Ren and Shobhit Verma and Anton Clarkson and Mingfei Yan and Brian Budge and Yajie Yan and Xiaqing Pan and June Yon and Yuyang Zou and Kimberly Leon and Nigel Carter and Jesus Briales and  Tyler Gillingham and  Elias Mueggler and Luis Pesqueira and Manolis Savva and Dhruv Batra and Hauke M. Strasdat and Renzo De Nardi and Michael Goesele and Steven Lovegrove and Richard Newcombe },\n  journal = {arXiv preprint arXiv:1906.05797},\n  year =    {2019}\n}\n```\n\n## Replica 数据集\n\n以下 18 个场景包含在此次初始发布中：\n\n![Replica Dataset](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Ffacebookresearch_Replica-Dataset_readme_c3098187eb10.png)\n\n![Replica Dataset](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Ffacebookresearch_Replica-Dataset_readme_6a9a398b139f.png)\n\n每个 Replica 包含以下资源：\n```\n├── glass.sur\n├── habitat\n    ├── mesh_semantic.ply \n    ├── mesh_semantic.navmesh\n    ├── info_semantic.json\n    ├── mesh_preseg_semantic.ply \n    ├── mesh_preseg_semantic.navmesh\n    ├── info_preseg_semantic.json\n    ├── replica_stage.stage_config.json\n    └── sorted_faces.bin \n├── mesh.ply\n├── preseg.bin\n├── preseg.json\n├── semantic.bin\n├── semantic.json\n└── textures\n    ├── 0-color-ptex.hdr\n    ├── 0-color-ptex.w\n    ├── 1-color-ptex.hdr\n    ├── 1-color-ptex.w\n    ├── ...\n    └── parameters.json\n```\n不同文件包含以下内容：\n- `glass.sur`: 玻璃和镜面表面的参数化。\n- `mesh.ply`: 带有顶点颜色的场景四边形网格。\n- `preseg.json` 和 `preseg.bin`: 场景中平面和非平面的预分割 (Presegmentation)。\n- `semantic.json` 和 `semantic.bin`: 场景的语义分割 (Semantic Segmentation)。\n- `textures`: 场景的高分辨率和高动态范围 (HDR) 纹理。\n- `habitat\u002Fmesh*semantic.ply`: 包含语义或预分割 (Presegmentation) 信息的四边形网格，适用于 AI Habitat。 \n- `habitat\u002Finfo*semantic.json`: 将相应 `mesh_*.ply` 中的实例 ID 映射到语义名称。\n- `habitat\u002Fmesh*semantic.navmesh`: AI Habitat 的导航网格 (Navmesh)。\n- `habitat\u002Freplica_stage.stage_config.json`: 定义 habitat-sim 场景级别参数的配置文件。\n- `habitat\u002Fsorted_faces.bin`: 包含 habitat-sim Ptex 渲染支持预处理几何数据的二进制文件。\n\n### 在 Mac OS 和 Linux 上下载\n确保已安装 `pigz`、`wget` 和 `unzip`：\n```\n# on Mac OS\nbrew install wget pigz unzip\n# on Ubuntu\nsudo apt-get install wget pigz unzip\n```\n要下载并解压数据集，请使用 `download.sh` 脚本：\n```\n.\u002Fdownload.sh \u002Fpath\u002Fto\u002Freplica_v1\n```\n\n### 在 Windows 上下载\n\n执行 `win_download.bat` 以下载 Replica。\n\n## Replica SDK\n\n### 设置\n在安装 [Pangolin](https:\u002F\u002Fgithub.com\u002Fstevenlovegrove\u002FPangolin) 的依赖项后，可以通过构建脚本编译 Replica SDK（软件开发工具包）：\n```\ngit submodule update --init\n.\u002Fbuild.sh\n```\n它需要安装 [Pangolin](https:\u002F\u002Fgithub.com\u002Fstevenlovegrove\u002FPangolin) 和 [Eigen](https:\u002F\u002Fgithub.com\u002Feigenteam\u002Feigen-git-mirror) 的依赖项。如果您希望使用无头模式 (Headless) 渲染器，请确保您已安装 libegl1-mesa-dev 包。\n\n### ReplicaViewer\n\nReplicaViewer 是一个用于探索 Replica 数据集的交互式用户界面 (UI)。 \n\n```\n.\u002Fbuild\u002Fbin\u002FReplicaViewer mesh.ply \u002Fpath\u002Fto\u002Fatlases [mirrorFile]\n```\n\n![ReplicaViewer](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Ffacebookresearch_Replica-Dataset_readme_49dd4ff82dd5.png)\n\n来自高动态范围 (HDR) 纹理的渲染曝光值可以在左上角进行调整。 \n\n### ReplicaRenderer\n\nReplicaRenderer 展示了如何为程序定义的轨迹从 Replica 中渲染图像，无需用户界面 (UI)。此可执行文件可以在服务器上无头模式 (Headless) 运行。 \n\n```\n.\u002Fbuild\u002Fbin\u002FReplicaRenderer mesh.ply textures glass.sur\n```\n\n## Replica 与 AI Habitat\n\n要在 AI Habitat 中使用 Replica，请查看 [https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fhabitat-sim](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fhabitat-sim) 处的 AI Habitat Sim。构建项目后，您可以启动测试查看器以验证一切是否正常：\n```\n.\u002Fbuild\u002Fviewer --dataset \u002FPATH\u002FTO\u002FREPLICA\u002Freplica.scene_dataset_config.json -- frl_apartment_0\n```\n\n## 团队 \n\nJulian Straub,  Thomas Whelan, Lingni Ma, Yufan Chen, Erik Wijmans, Simon Green, Jakob J. Engel, Raul Mur-Artal, Carl Ren, Shobhit Verma, Anton Clarkson, Mingfei Yan, Brian Budge, Yajie Yan, Xiaqing Pan, June Yon, Yuyang Zou, Kimberly Leon, Nigel Carter, Jesus Briales,  Tyler Gillingham Elias Mueggler, Luis Pesqueira, Manolis Savva, Dhruv Batra, Hauke M. Strasdat, Renzo De Nardi, Michael Goesele, Steven Lovegrove, and Richard Newcombe.\n\n## 联系方式\n\n[Julian.Straub@oculus.com](Julian.Straub@oculus.com)\n\n## 致谢\n\nReplica 数据集的实现离不开 Matthew Banks, Christopher Dotson, Rashad Barber, Justin Blosch, Ethan Henderson, Kelley Greene, Michael Thot, Matthew Winterscheid, Robert Johnston, Abhijit Kulkarni, Robert Meeker, Jamie Palacios, Tony Phan, Tim Petrvalsky, Sayed Farhad Sadat, Manuel Santana, Suruj Singh, Swati Agrawal, 和 Hannah Woolums 的辛勤工作和贡献。\n\n## 许可证\n\n有关详细信息，请参阅 LICENSE 文件。","# Replica Dataset 快速上手指南\n\n**Replica Dataset** 是一个高质量室内空间重建数据集，提供干净的稠密几何、高分辨率 HDR 纹理、玻璃\u002F镜面信息以及语义分割数据。该工具集支持通过 `ReplicaViewer` 进行可视化检查，通过 `ReplicaRenderer` 进行无头渲染，并可直接用于 **AI Habitat** 框架进行智能体训练。\n\n---\n\n## 环境准备\n\n在开始之前，请确保您的系统满足以下要求：\n\n*   **操作系统**: Mac OS, Linux (Ubuntu), Windows\n*   **下载工具**: `wget`, `pigz`, `unzip`\n*   **编译依赖**:\n    *   [Pangolin](https:\u002F\u002Fgithub.com\u002Fstevenlovegrove\u002FPangolin)\n    *   [Eigen](https:\u002F\u002Fgithub.com\u002Feigenteam\u002Feigen-git-mirror)\n    *   *(可选)* `libegl1-mesa-dev`：若需使用无头模式渲染器 (`ReplicaRenderer`)，请确保安装此包。\n\n---\n\n## 安装步骤\n\n### 1. 下载数据集\n\n根据操作系统选择对应的下载方式：\n\n**Mac OS 和 Linux**\n首先安装必要的下载工具：\n```bash\n# Mac OS\nbrew install wget pigz unzip\n# Ubuntu\nsudo apt-get install wget pigz unzip\n```\n然后运行脚本下载并解压数据集（请将路径替换为实际目标目录）：\n```bash\n.\u002Fdownload.sh \u002Fpath\u002Fto\u002Freplica_v1\n```\n\n**Windows**\n直接执行提供的批处理文件：\n```cmd\nwin_download.bat\n```\n\n### 2. 编译 SDK\n进入项目根目录，初始化子模块并编译：\n```bash\ngit submodule update --init\n.\u002Fbuild.sh\n```\n\n---\n\n## 基本使用\n\n### 1. 交互式可视化 (ReplicaViewer)\n用于探索数据集场景的交互界面。\n```bash\n.\u002Fbuild\u002Fbin\u002FReplicaViewer mesh.ply \u002Fpath\u002Fto\u002Fatlases [mirrorFile]\n```\n*注：可在左上角调整 HDR 纹理的曝光值。*\n\n### 2. 程序化渲染 (ReplicaRenderer)\n无需 UI 即可从场景中渲染图像，支持在无头服务器（Headless Server）上运行。\n```bash\n.\u002Fbuild\u002Fbin\u002FReplicaRenderer mesh.ply textures glass.sur\n```\n\n### 3. 集成 AI Habitat\n如需在 AI Habitat 中使用，请克隆 [habitat-sim](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fhabitat-sim)。构建完成后，可通过以下命令验证场景配置：\n```bash\n.\u002Fbuild\u002Fviewer --dataset \u002FPATH\u002FTO\u002FREPLICA\u002Freplica.scene_dataset_config.json -- frl_apartment_0\n```\n\n---\n\n> **引用说明**：若在研究中使用本数据集，请引用技术报告：\n> ```bibtex\n> @article{replica19arxiv,\n>   title =   {The {R}eplica Dataset: A Digital Replica of Indoor Spaces},\n>   author =  {Julian Straub et al.},\n>   journal = {arXiv preprint arXiv:1906.05797},\n>   year =    {2019}\n> }\n> ```","某机器人研发团队正在开发家庭服务机器人的自主导航系统，急需大量逼真且标注完善的室内仿真环境进行强化学习训练。\n\n### 没有 Replica-Dataset 时\n- 团队需手动建模或激光扫描真实房间，耗时数月且几何精度难以保证，无法满足大规模测试需求\n- 缺乏带语义标注和玻璃反射信息的场景，导致智能体无法理解复杂物理属性，容易在透明物体上碰撞\n- 数据格式不统一，需编写大量转换代码才能接入 AI Habitat 等主流仿真框架，工程维护成本高\n- 纹理分辨率低，仿真环境与现实世界差距大，导致训练好的模型在实际部署时泛化能力差\n\n### 使用 Replica-Dataset 后\n- 直接加载 18 个高精度室内场景，包含稠密几何与 HDR 纹理，开箱即用，立即开始实验\n- 内置语义分割、实例分割及玻璃镜面参数，无需额外标注即可训练智能体的感知与避障能力\n- 提供 AI Habitat 专用导出格式，无缝集成到现有仿真 pipeline 中，省去了繁琐的数据预处理步骤\n- 大幅缩短环境构建周期，从数周缩减至数小时，显著提升算法迭代效率并降低硬件成本\n\n通过提供标准化的高质量室内数字孪生数据，Replica-Dataset 极大降低了具身智能仿真环境的构建门槛。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Ffacebookresearch_Replica-Dataset_11418741.png","facebookresearch","Meta Research","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Ffacebookresearch_449342bd.png","",null,"https:\u002F\u002Fopensource.fb.com","https:\u002F\u002Fgithub.com\u002Ffacebookresearch",[83,87,91,95,99],{"name":84,"color":85,"percentage":86},"C++","#f34b7d",81.6,{"name":88,"color":89,"percentage":90},"GLSL","#5686a5",10.9,{"name":92,"color":93,"percentage":94},"CMake","#DA3434",4.8,{"name":96,"color":97,"percentage":98},"Shell","#89e051",1.3,{"name":100,"color":101,"percentage":98},"Batchfile","#C1F12E",1240,108,"2026-04-05T13:46:38","NOASSERTION","Linux, macOS, Windows","未说明",{"notes":109,"python":107,"dependencies":110},"支持在 Mac OS、Linux 和 Windows 系统上下载数据集。SDK 编译依赖 Pangolin 和 Eigen 库。若需在服务器上无头运行渲染器，Linux 需安装 libegl1-mesa-dev。数据集已导出为 AI Habitat 格式，可直接用于该框架的训练任务。使用时请引用相关技术报告。",[111,112,113,114,115,116],"Pangolin","Eigen","libegl1-mesa-dev","pigz","wget","unzip",[46,15],"2026-03-27T02:49:30.150509","2026-04-06T08:18:25.910670",[121,126,131,136,141,146],{"id":122,"question_zh":123,"answer_zh":124,"source_url":125},2509,"下载 Replica 数据集时遇到 'unpigz: command not found' 错误如何解决？","如果在服务器上无法获取 root 权限，可以通过 conda 安装 pigz：`conda install -c conda-forge pigz`。如果是 Mac OS 用户，请先安装 Homebrew，然后运行 `brew install pigz`。","https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002FReplica-Dataset\u002Fissues\u002F12",{"id":127,"question_zh":128,"answer_zh":129,"source_url":130},2510,"运行 ReplicaViewer 时报错 'Assert failed: pangolin::FileExists(atlasFolder)' 是什么原因？","这是因为纹理文件夹路径参数不正确。`textures` 参数必须指向包含 `.hdr` 文件的文件夹。正确的命令格式应为：`.\u002Fbuild\u002FReplicaSDK\u002FReplicaViewer .\u002Fdata\u002Fapartment_0\u002Fmesh.ply .\u002Fdata\u002Fapartment_0\u002Ftextures\u002F .\u002Fdata\u002Fapartment_0\u002Fglass.sur`。","https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002FReplica-Dataset\u002Fissues\u002F72",{"id":132,"question_zh":133,"answer_zh":134,"source_url":135},2511,"在 Habitat-Sim 中加载 Replica 场景时出现 Shader 链接失败（error C6022）怎么办？","尝试显式指定语义网格路径，例如传入 `\u002Fpath_to_replicadataset\u002Fhabitat\u002Fmesh_semantic.ply`。注意：仅修改路径可能导致光照和反射效果消失，更完善的解决方案是确保几何着色器与顶点和片段着色器正确链接。","https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002FReplica-Dataset\u002Fissues\u002F90",{"id":137,"question_zh":138,"answer_zh":139,"source_url":140},2512,"Habitat-Sim 与 ReplicaSDK 之间的坐标系方向有何不同？","Habitat 的相机向下看 -Z 轴，而 ReplicaSDK 的相机向下看 +Z 轴。这种差异主要影响代理的旋转而非位置。需要使用变换矩阵 `T_replicac_hsimc`（将 -UnitZ 旋转到 +UnitZ）来进行坐标转换。","https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002FReplica-Dataset\u002Fissues\u002F29",{"id":142,"question_zh":143,"answer_zh":144,"source_url":145},2513,"运行 ReplicaRenderer 时发生段错误（Segmentation fault）该如何排查？","首先尝试安装缺失的系统依赖库，如 `sudo apt install libegl1-mesa-dev`。如果问题依旧存在，建议切换到 Docker 环境运行，以避免本地系统配置导致的兼容性问题。","https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002FReplica-Dataset\u002Fissues\u002F8",{"id":147,"question_zh":148,"answer_zh":149,"source_url":150},2514,"在 Habitat 中使用 Replica 数据集时，是否需要将 .ply 文件转换为 .glb 格式？","不需要。数据集中应为每个环境提供了对应的 .ply 文件（如 mesh_semantic.ply），可以直接使用这些 .ply 文件生成训练数据，无需通过 assimp 转换为 .glb。","https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002FReplica-Dataset\u002Fissues\u002F63",[152],{"id":153,"version":154,"summary_zh":155,"released_at":156},111711,"v1.0","The Replica dataset as described in https:\u002F\u002Farxiv.org\u002Fpdf\u002F1906.05797.pdf","2019-06-14T15:30:52"]