[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-isaac-sim--IsaacLab":3,"tool-isaac-sim--IsaacLab":65},[4,16,31,40,48,57],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":15},4721,"markitdown","microsoft\u002Fmarkitdown","MarkItDown 是一款由微软 AutoGen 团队打造的轻量级 Python 工具，专为将各类文件高效转换为 Markdown 格式而设计。它支持 PDF、Word、Excel、PPT、图片（含 OCR）、音频（含语音转录）、HTML 乃至 YouTube 链接等多种格式的解析，能够精准提取文档中的标题、列表、表格和链接等关键结构信息。\n\n在人工智能应用日益普及的今天，大语言模型（LLM）虽擅长处理文本，却难以直接读取复杂的二进制办公文档。MarkItDown 恰好解决了这一痛点，它将非结构化或半结构化的文件转化为模型“原生理解”且 Token 效率极高的 Markdown 格式，成为连接本地文件与 AI 分析 pipeline 的理想桥梁。此外，它还提供了 MCP（模型上下文协议）服务器，可无缝集成到 Claude Desktop 等 LLM 应用中。\n\n这款工具特别适合开发者、数据科学家及 AI 研究人员使用，尤其是那些需要构建文档检索增强生成（RAG）系统、进行批量文本分析或希望让 AI 助手直接“阅读”本地文件的用户。虽然生成的内容也具备一定可读性，但其核心优势在于为机器",93400,2,"2026-04-06T19:52:38",[13,14],"插件","开发框架","ready",{"id":17,"name":18,"github_repo":19,"description_zh":20,"stars":21,"difficulty_score":10,"last_commit_at":22,"category_tags":23,"status":15},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。",85013,"2026-04-06T11:09:19",[24,25,26,13,27,28,29,14,30],"图像","数据工具","视频","Agent","其他","语言模型","音频",{"id":32,"name":33,"github_repo":34,"description_zh":35,"stars":36,"difficulty_score":37,"last_commit_at":38,"category_tags":39,"status":15},3128,"ragflow","infiniflow\u002Fragflow","RAGFlow 是一款领先的开源检索增强生成（RAG）引擎，旨在为大语言模型构建更精准、可靠的上下文层。它巧妙地将前沿的 RAG 技术与智能体（Agent）能力相结合，不仅支持从各类文档中高效提取知识，还能让模型基于这些知识进行逻辑推理和任务执行。\n\n在大模型应用中，幻觉问题和知识滞后是常见痛点。RAGFlow 通过深度解析复杂文档结构（如表格、图表及混合排版），显著提升了信息检索的准确度，从而有效减少模型“胡编乱造”的现象，确保回答既有据可依又具备时效性。其内置的智能体机制更进一步，使系统不仅能回答问题，还能自主规划步骤解决复杂问题。\n\n这款工具特别适合开发者、企业技术团队以及 AI 研究人员使用。无论是希望快速搭建私有知识库问答系统，还是致力于探索大模型在垂直领域落地的创新者，都能从中受益。RAGFlow 提供了可视化的工作流编排界面和灵活的 API 接口，既降低了非算法背景用户的上手门槛，也满足了专业开发者对系统深度定制的需求。作为基于 Apache 2.0 协议开源的项目，它正成为连接通用大模型与行业专有知识之间的重要桥梁。",77062,3,"2026-04-04T04:44:48",[27,24,14,29,28],{"id":41,"name":42,"github_repo":43,"description_zh":44,"stars":45,"difficulty_score":37,"last_commit_at":46,"category_tags":47,"status":15},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。",74963,"2026-04-06T11:16:39",[29,24,14,28],{"id":49,"name":50,"github_repo":51,"description_zh":52,"stars":53,"difficulty_score":54,"last_commit_at":55,"category_tags":56,"status":15},3215,"awesome-machine-learning","josephmisiti\u002Fawesome-machine-learning","awesome-machine-learning 是一份精心整理的机器学习资源清单，汇集了全球优秀的机器学习框架、库和软件工具。面对机器学习领域技术迭代快、资源分散且难以甄选的痛点，这份清单按编程语言（如 Python、C++、Go 等）和应用场景（如计算机视觉、自然语言处理、深度学习等）进行了系统化分类，帮助使用者快速定位高质量项目。\n\n它特别适合开发者、数据科学家及研究人员使用。无论是初学者寻找入门库，还是资深工程师对比不同语言的技术选型，都能从中获得极具价值的参考。此外，清单还延伸提供了免费书籍、在线课程、行业会议、技术博客及线下聚会等丰富资源，构建了从学习到实践的全链路支持体系。\n\n其独特亮点在于严格的维护标准：明确标记已停止维护或长期未更新的项目，确保推荐内容的时效性与可靠性。作为机器学习领域的“导航图”，awesome-machine-learning 以开源协作的方式持续更新，旨在降低技术探索门槛，让每一位从业者都能高效地站在巨人的肩膀上创新。",72149,1,"2026-04-03T21:50:24",[14,28],{"id":58,"name":59,"github_repo":60,"description_zh":61,"stars":62,"difficulty_score":37,"last_commit_at":63,"category_tags":64,"status":15},2181,"OpenHands","OpenHands\u002FOpenHands","OpenHands 是一个专注于 AI 驱动开发的开源平台，旨在让智能体（Agent）像人类开发者一样理解、编写和调试代码。它解决了传统编程中重复性劳动多、环境配置复杂以及人机协作效率低等痛点，通过自动化流程显著提升开发速度。\n\n无论是希望提升编码效率的软件工程师、探索智能体技术的研究人员，还是需要快速原型验证的技术团队，都能从中受益。OpenHands 提供了灵活多样的使用方式：既可以通过命令行（CLI）或本地图形界面在个人电脑上轻松上手，体验类似 Devin 的流畅交互；也能利用其强大的 Python SDK 自定义智能体逻辑，甚至在云端大规模部署上千个智能体并行工作。\n\n其核心技术亮点在于模块化的软件智能体 SDK，这不仅构成了平台的引擎，还支持高度可组合的开发模式。此外，OpenHands 在 SWE-bench 基准测试中取得了 77.6% 的优异成绩，证明了其解决真实世界软件工程问题的能力。平台还具备完善的企业级功能，支持与 Slack、Jira 等工具集成，并提供细粒度的权限管理，适合从个人开发者到大型企业的各类用户场景。",70665,"2026-04-06T11:28:43",[29,27,14,13],{"id":66,"github_repo":67,"name":68,"description_en":69,"description_zh":70,"ai_summary_zh":70,"readme_en":71,"readme_zh":72,"quickstart_zh":73,"use_case_zh":74,"hero_image_url":75,"owner_login":76,"owner_name":77,"owner_avatar_url":78,"owner_bio":79,"owner_company":80,"owner_location":80,"owner_email":80,"owner_twitter":80,"owner_website":80,"owner_url":81,"languages":82,"stars":106,"forks":107,"last_commit_at":108,"license":109,"difficulty_score":110,"env_os":111,"env_gpu":112,"env_ram":113,"env_deps":114,"category_tags":124,"github_topics":125,"view_count":10,"oss_zip_url":80,"oss_zip_packed_at":80,"status":15,"created_at":129,"updated_at":130,"faqs":131,"releases":159},4626,"isaac-sim\u002FIsaacLab","IsaacLab","Unified framework for robot learning built on NVIDIA Isaac Sim","Isaac Lab 是一个基于 NVIDIA Isaac Sim 构建的开源统一框架，旨在简化和加速机器人学习的研究工作流。它专为强化学习、模仿学习和运动规划等任务设计，通过提供高度逼真的物理引擎和传感器模拟，有效解决了机器人算法从虚拟仿真到真实世界部署（Sim-to-Real）过程中面临的效率低、精度差及环境复现难等核心痛点。\n\n这款工具非常适合机器人领域的研究人员、算法开发者以及高校师生使用。无论是需要快速验证新算法的学术探索，还是进行大规模数据训练的工程落地，Isaac Lab 都能提供强有力的支持。其独特的技术亮点在于全面的 GPU 加速能力，能够并行运行复杂的计算任务，大幅缩短迭代周期；同时内置了超过 16 种主流机器人模型（如机械臂、四足机器人和人形机器人）及 30 多个开箱即用的训练环境，并支持 RTX 渲染的高精度相机、激光雷达等多种传感器仿真。此外，它还兼容多种流行的强化学习库，既支持本地运行也能扩展至云端，为不同规模的项目提供了极大的灵活性。","![Isaac Lab](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fisaac-sim_IsaacLab_readme_a7952be5e9c3.jpg)\n\n---\n\n# Isaac Lab\n\n[![IsaacSim](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FIsaacSim-5.1.0-silver.svg)](https:\u002F\u002Fdocs.isaacsim.omniverse.nvidia.com\u002Flatest\u002Findex.html)\n[![Python](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpython-3.11-blue.svg)](https:\u002F\u002Fdocs.python.org\u002F3\u002Fwhatsnew\u002F3.11.html)\n[![Linux platform](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fplatform-linux--64-orange.svg)](https:\u002F\u002Freleases.ubuntu.com\u002F22.04\u002F)\n[![Windows platform](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fplatform-windows--64-orange.svg)](https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002F)\n[![pre-commit](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002Fisaac-sim\u002FIsaacLab\u002Fpre-commit.yaml?logo=pre-commit&logoColor=white&label=pre-commit&color=brightgreen)](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Factions\u002Fworkflows\u002Fpre-commit.yaml)\n[![docs status](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002Fisaac-sim\u002FIsaacLab\u002Fdocs.yaml?label=docs&color=brightgreen)](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Factions\u002Fworkflows\u002Fdocs.yaml)\n[![License](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-BSD--3-yellow.svg)](https:\u002F\u002Fopensource.org\u002Flicenses\u002FBSD-3-Clause)\n[![License](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-Apache--2.0-yellow.svg)](https:\u002F\u002Fopensource.org\u002Flicense\u002Fapache-2-0)\n\n\n**Isaac Lab** is a GPU-accelerated, open-source framework designed to unify and simplify robotics research workflows,\nsuch as reinforcement learning, imitation learning, and motion planning. Built on [NVIDIA Isaac Sim](https:\u002F\u002Fdocs.isaacsim.omniverse.nvidia.com\u002Flatest\u002Findex.html),\nit combines fast and accurate physics and sensor simulation, making it an ideal choice for sim-to-real\ntransfer in robotics.\n\nIsaac Lab provides developers with a range of essential features for accurate sensor simulation, such as RTX-based\ncameras, LIDAR, or contact sensors. The framework's GPU acceleration enables users to run complex simulations and\ncomputations faster, which is key for iterative processes like reinforcement learning and data-intensive tasks.\nMoreover, Isaac Lab can run locally or be distributed across the cloud, offering flexibility for large-scale deployments.\n\nA detailed description of Isaac Lab can be found in our [arXiv paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2511.04831).\n\n## Key Features\n\nIsaac Lab offers a comprehensive set of tools and environments designed to facilitate robot learning:\n\n- **Robots**: A diverse collection of robots, from manipulators, quadrupeds, to humanoids, with more than 16 commonly available models.\n- **Environments**: Ready-to-train implementations of more than 30 environments, which can be trained with popular reinforcement learning frameworks such as RSL RL, SKRL, RL Games, or Stable Baselines. We also support multi-agent reinforcement learning.\n- **Physics**: Rigid bodies, articulated systems, deformable objects\n- **Sensors**: RGB\u002Fdepth\u002Fsegmentation cameras, camera annotations, IMU, contact sensors, ray casters.\n\n\n## Getting Started\n\n### Documentation\n\nOur [documentation page](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab) provides everything you need to get started, including\ndetailed tutorials and step-by-step guides. Follow these links to learn more about:\n\n- [Installation steps](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fmain\u002Fsource\u002Fsetup\u002Finstallation\u002Findex.html#local-installation)\n- [Reinforcement learning](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fmain\u002Fsource\u002Foverview\u002Freinforcement-learning\u002Frl_existing_scripts.html)\n- [Tutorials](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fmain\u002Fsource\u002Ftutorials\u002Findex.html)\n- [Available environments](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fmain\u002Fsource\u002Foverview\u002Fenvironments.html)\n\n\n## Isaac Sim Version Dependency\n\nIsaac Lab is built on top of Isaac Sim and requires specific versions of Isaac Sim that are compatible with each\nrelease of Isaac Lab. Below, we outline the recent Isaac Lab releases and GitHub branches and their corresponding\ndependency versions for Isaac Sim.\n\n| Isaac Lab Version             | Isaac Sim Version         |\n| ----------------------------- | ------------------------- |\n| `main` branch                 | Isaac Sim 4.5 \u002F 5.0 \u002F 5.1 |\n| `v2.3.X`                      | Isaac Sim 4.5 \u002F 5.0 \u002F 5.1 |\n| `v2.2.X`                      | Isaac Sim 4.5 \u002F 5.0       |\n| `v2.1.X`                      | Isaac Sim 4.5             |\n| `v2.0.X`                      | Isaac Sim 4.5             |\n\n\n## Contributing to Isaac Lab\n\nWe wholeheartedly welcome contributions from the community to make this framework mature and useful for everyone.\nThese may happen as bug reports, feature requests, or code contributions. For details, please check our\n[contribution guidelines](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fmain\u002Fsource\u002Frefs\u002Fcontributing.html).\n\n## Show & Tell: Share Your Inspiration\n\nWe encourage you to utilize our [Show & Tell](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fdiscussions\u002Fcategories\u002Fshow-and-tell)\narea in the `Discussions` section of this repository. This space is designed for you to:\n\n* Share the tutorials you've created\n* Showcase your learning content\n* Present exciting projects you've developed\n\nBy sharing your work, you'll inspire others and contribute to the collective knowledge\nof our community. Your contributions can spark new ideas and collaborations, fostering\ninnovation in robotics and simulation.\n\n## Troubleshooting\n\nPlease see the [troubleshooting](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fmain\u002Fsource\u002Frefs\u002Ftroubleshooting.html) section for\ncommon fixes or [submit an issue](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fissues).\n\nFor issues related to Isaac Sim, we recommend checking its [documentation](https:\u002F\u002Fdocs.isaacsim.omniverse.nvidia.com\u002Flatest\u002Findex.html)\nor opening a question on its [forums](https:\u002F\u002Fforums.developer.nvidia.com\u002Fc\u002Fagx-autonomous-machines\u002Fisaac\u002F67).\n\n## Support\n\n* Please use GitHub [Discussions](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fdiscussions) for discussing ideas,\n  asking questions, and requests for new features.\n* Github [Issues](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fissues) should only be used to track executable pieces of\n  work with a definite scope and a clear deliverable. These can be fixing bugs, documentation issues, new features,\n  or general updates.\n\n## Connect with the NVIDIA Omniverse Community\n\nDo you have a project or resource you'd like to share more widely? We'd love to hear from you!\nReach out to the NVIDIA Omniverse Community team at OmniverseCommunity@nvidia.com to explore opportunities\nto spotlight your work.\n\nYou can also join the conversation on the [Omniverse Discord](https:\u002F\u002Fdiscord.com\u002Finvite\u002Fnvidiaomniverse) to\nconnect with other developers, share your projects, and help grow a vibrant, collaborative ecosystem\nwhere creativity and technology intersect. Your contributions can make a meaningful impact on the Isaac Lab\ncommunity and beyond!\n\n## License\n\nThe Isaac Lab framework is released under [BSD-3 License](LICENSE). The `isaaclab_mimic` extension and its\ncorresponding standalone scripts are released under [Apache 2.0](LICENSE-mimic). The license files of its\ndependencies and assets are present in the [`docs\u002Flicenses`](docs\u002Flicenses) directory.\n\nNote that Isaac Lab requires Isaac Sim, which includes components under proprietary licensing terms. Please see the [Isaac Sim license](docs\u002Flicenses\u002Fdependencies\u002Fisaacsim-license.txt) for information on Isaac Sim licensing.\n\nNote that the `isaaclab_mimic` extension requires cuRobo, which has proprietary licensing terms that can be found in [`docs\u002Flicenses\u002Fdependencies\u002FcuRobo-license.txt`](docs\u002Flicenses\u002Fdependencies\u002FcuRobo-license.txt).\n\n\n## Citation\n\nIf you use Isaac Lab in your research, please cite the technical report:\n\n```\n@article{mittal2025isaaclab,\n  title={Isaac Lab: A GPU-Accelerated Simulation Framework for Multi-Modal Robot Learning},\n  author={Mayank Mittal and Pascal Roth and James Tigue and Antoine Richard and Octi Zhang and Peter Du and Antonio Serrano-Muñoz and Xinjie Yao and René Zurbrügg and Nikita Rudin and Lukasz Wawrzyniak and Milad Rakhsha and Alain Denzler and Eric Heiden and Ales Borovicka and Ossama Ahmed and Iretiayo Akinola and Abrar Anwar and Mark T. Carlson and Ji Yuan Feng and Animesh Garg and Renato Gasoto and Lionel Gulich and Yijie Guo and M. Gussert and Alex Hansen and Mihir Kulkarni and Chenran Li and Wei Liu and Viktor Makoviychuk and Grzegorz Malczyk and Hammad Mazhar and Masoud Moghani and Adithyavairavan Murali and Michael Noseworthy and Alexander Poddubny and Nathan Ratliff and Welf Rehberg and Clemens Schwarke and Ritvik Singh and James Latham Smith and Bingjie Tang and Ruchik Thaker and Matthew Trepte and Karl Van Wyk and Fangzhou Yu and Alex Millane and Vikram Ramasamy and Remo Steiner and Sangeeta Subramanian and Clemens Volk and CY Chen and Neel Jawale and Ashwin Varghese Kuruttukulam and Michael A. Lin and Ajay Mandlekar and Karsten Patzwaldt and John Welsh and Huihua Zhao and Fatima Anes and Jean-Francois Lafleche and Nicolas Moënne-Loccoz and Soowan Park and Rob Stepinski and Dirk Van Gelder and Chris Amevor and Jan Carius and Jumyung Chang and Anka He Chen and Pablo de Heras Ciechomski and Gilles Daviet and Mohammad Mohajerani and Julia von Muralt and Viktor Reutskyy and Michael Sauter and Simon Schirm and Eric L. Shi and Pierre Terdiman and Kenny Vilella and Tobias Widmer and Gordon Yeoman and Tiffany Chen and Sergey Grizan and Cathy Li and Lotus Li and Connor Smith and Rafael Wiltz and Kostas Alexis and Yan Chang and David Chu and Linxi \"Jim\" Fan and Farbod Farshidian and Ankur Handa and Spencer Huang and Marco Hutter and Yashraj Narang and Soha Pouya and Shiwei Sheng and Yuke Zhu and Miles Macklin and Adam Moravanszky and Philipp Reist and Yunrong Guo and David Hoeller and Gavriel State},\n  journal={arXiv preprint arXiv:2511.04831},\n  year={2025},\n  url={https:\u002F\u002Farxiv.org\u002Fabs\u002F2511.04831}\n}\n```\n\n## Acknowledgement\n\nIsaac Lab development initiated from the [Orbit](https:\u002F\u002Fisaac-orbit.github.io\u002F) framework.\nWe gratefully acknowledge the authors of Orbit for their foundational contributions.\n","![艾萨克实验室](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fisaac-sim_IsaacLab_readme_a7952be5e9c3.jpg)\n\n---\n\n# 艾萨克实验室\n\n[![IsaacSim](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FIsaacSim-5.1.0-silver.svg)](https:\u002F\u002Fdocs.isaacsim.omniverse.nvidia.com\u002Flatest\u002Findex.html)\n[![Python](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpython-3.11-blue.svg)](https:\u002F\u002Fdocs.python.org\u002F3\u002Fwhatsnew\u002F3.11.html)\n[![Linux平台](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fplatform-linux--64-orange.svg)](https:\u002F\u002Freleases.ubuntu.com\u002F22.04\u002F)\n[![Windows平台](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fplatform-windows--64-orange.svg)](https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002F)\n[![pre-commit](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002Fisaac-sim\u002FIsaacLab\u002Fpre-commit.yaml?logo=pre-commit&logoColor=white&label=pre-commit&color=brightgreen)](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Factions\u002Fworkflows\u002Fpre-commit.yaml)\n[![文档状态](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002Fisaac-sim\u002FIsaacLab\u002Fdocs.yaml?label=docs&color=brightgreen)](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Factions\u002Fworkflows\u002Fdocs.yaml)\n[![许可证](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-BSD--3-yellow.svg)](https:\u002F\u002Fopensource.org\u002Flicenses\u002FBSD-3-Clause)\n[![许可证](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-Apache--2.0-yellow.svg)](https:\u002F\u002Fopensource.org\u002Flicense\u002Fapache-2-0)\n\n\n**艾萨克实验室** 是一个基于 GPU 加速的开源框架，旨在统一并简化机器人研究工作流，例如强化学习、模仿学习和运动规划。它构建于 [NVIDIA Isaac Sim](https:\u002F\u002Fdocs.isaacsim.omniverse.nvidia.com\u002Flatest\u002Findex.html) 之上，结合了快速且精确的物理与传感器仿真，使其成为机器人领域从仿真到现实迁移的理想选择。\n\n艾萨克实验室为开发者提供了多种用于精确传感器仿真的关键功能，例如基于 RTX 的相机、激光雷达或接触传感器。该框架的 GPU 加速能力使用户能够更快地运行复杂的仿真和计算，这对于强化学习等迭代过程以及数据密集型任务至关重要。此外，艾萨克实验室既可以在本地运行，也可以在云端分布式部署，从而为大规模应用提供灵活性。\n\n有关艾萨克实验室的详细描述，请参阅我们的 [arXiv 论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2511.04831)。\n\n## 核心特性\n\n艾萨克实验室提供了一套全面的工具和环境，专为促进机器人学习而设计：\n\n- **机器人**：多样化的机器人集合，涵盖机械臂、四足机器人和人形机器人等多种类型，包含超过 16 种常用模型。\n- **环境**：超过 30 种即用型训练环境，可与 RSL RL、SKRL、RL Games 或 Stable Baselines 等主流强化学习框架配合使用。我们还支持多智能体强化学习。\n- **物理引擎**：刚体、关节系统和可变形物体。\n- **传感器**：RGB\u002F深度\u002F分割相机、相机标注、IMU、接触传感器、射线投射器。\n\n\n## 快速入门\n\n### 文档\n\n我们的 [文档页面](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab) 提供了您开始所需的一切，包括详细的教程和分步指南。请通过以下链接了解更多：\n\n- [安装步骤](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fmain\u002Fsource\u002Fsetup\u002Finstallation\u002Findex.html#local-installation)\n- [强化学习](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fmain\u002Fsource\u002Foverview\u002Freinforcement-learning\u002Frl_existing_scripts.html)\n- [教程](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fmain\u002Fsource\u002Ftutorials\u002Findex.html)\n- [可用环境](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fmain\u002Fsource\u002Foverview\u002Fenvironments.html)\n\n\n## Isaac Sim 版本依赖\n\n艾萨克实验室构建于 Isaac Sim 之上，因此需要与每个艾萨克实验室版本兼容的特定 Isaac Sim 版本。以下是近期艾萨克实验室发布版本及其对应的 GitHub 分支，以及它们对 Isaac Sim 的版本依赖关系。\n\n| 艾萨克实验室版本             | Isaac Sim 版本         |\n| ----------------------------- | ------------------------- |\n| `main` 分支                 | Isaac Sim 4.5 \u002F 5.0 \u002F 5.1 |\n| `v2.3.X`                      | Isaac Sim 4.5 \u002F 5.0 \u002F 5.1 |\n| `v2.2.X`                      | Isaac Sim 4.5 \u002F 5.0       |\n| `v2.1.X`                      | Isaac Sim 4.5             |\n| `v2.0.X`                      | Isaac Sim 4.5             |\n\n\n## 参与艾萨克实验室开发\n\n我们诚挚欢迎社区贡献，以使这一框架更加成熟并惠及所有人。这些贡献可以是漏洞报告、功能请求或代码提交。有关详情，请查阅我们的 [贡献指南](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fmain\u002Fsource\u002Frefs\u002Fcontributing.html)。\n\n## 展示与分享：分享您的灵感\n\n我们鼓励您利用此仓库“讨论”版块中的 [展示与分享](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fdiscussions\u002Fcategories\u002Fshow-and-tell) 区域。该空间专为您设计，用于：\n\n* 分享您创建的教程\n* 展示您的学习内容\n* 呈现您开发的精彩项目\n\n通过分享您的作品，您将激励他人，并为社区的集体知识库做出贡献。您的贡献能够激发新的想法和合作，推动机器人与仿真领域的创新。\n\n## 故障排除\n\n请参阅 [故障排除](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fmain\u002Fsource\u002Frefs\u002Ftroubleshooting.html) 部分，获取常见问题的解决方法，或 [提交问题](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fissues)。\n\n如遇与 Isaac Sim 相关的问题，建议查阅其 [文档](https:\u002F\u002Fdocs.isaacsim.omniverse.nvidia.com\u002Flatest\u002Findex.html) 或在其 [论坛](https:\u002F\u002Fforums.developer.nvidia.com\u002Fc\u002Fagx-autonomous-machines\u002Fisaac\u002F67) 上提问。\n\n## 支持\n\n* 请使用 GitHub 的 [讨论区](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fdiscussions) 来交流想法、提出问题以及请求新功能。\n* GitHub 的 [问题](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fissues) 应仅用于跟踪具有明确范围和清晰交付成果的具体工作，例如修复 bug、文档问题、新增功能或常规更新。\n\n## 与 NVIDIA Omniverse 社区建立联系\n\n您是否有想要更广泛分享的项目或资源？我们非常期待您的来信！请发送邮件至 OmniverseCommunity@nvidia.com，与 NVIDIA Omniverse 社区团队探讨如何突出您的工作。\n\n您还可以加入 [Omniverse Discord](https:\u002F\u002Fdiscord.com\u002Finvite\u002Fnvidiaomniverse) 社区，与其他开发者交流、分享您的项目，并共同打造一个充满活力、协作共生的生态系统，在这里创意与技术交汇融合。您的贡献将对艾萨克实验室社区乃至更广阔的领域产生深远影响！\n\n## 许可证\n\nIsaac Lab 框架根据 [BSD-3 许可证](LICENSE) 发布。`isaaclab_mimic` 扩展及其相应的独立脚本则根据 [Apache 2.0](LICENSE-mimic) 发布。其依赖项和资产的许可证文件位于 [`docs\u002Flicenses`](docs\u002Flicenses) 目录中。\n\n请注意，Isaac Lab 需要 Isaac Sim，而 Isaac Sim 包含受专有许可条款约束的组件。有关 Isaac Sim 的许可信息，请参阅 [Isaac Sim 许可证](docs\u002Flicenses\u002Fdependencies\u002Fisaacsim-license.txt)。\n\n此外，`isaaclab_mimic` 扩展需要 cuRobo，cuRobo 也采用专有许可条款，相关许可信息可在 [`docs\u002Flicenses\u002Fdependencies\u002FcuRobo-license.txt`](docs\u002Flicenses\u002Fdependencies\u002FcuRobo-license.txt) 中找到。\n\n\n## 引用\n\n如果您在研究中使用了 Isaac Lab，请引用以下技术报告：\n\n```\n@article{mittal2025isaaclab,\n  title={Isaac Lab: A GPU-Accelerated Simulation Framework for Multi-Modal Robot Learning},\n  author={Mayank Mittal and Pascal Roth and James Tigue and Antoine Richard and Octi Zhang and Peter Du and Antonio Serrano-Muñoz and Xinjie Yao and René Zurbrügg and Nikita Rudin and Lukasz Wawrzyniak and Milad Rakhsha and Alain Denzler and Eric Heiden and Ales Borovicka and Ossama Ahmed and Iretiayo Akinola and Abrar Anwar and Mark T. Carlson and Ji Yuan Feng and Animesh Garg and Renato Gasoto and Lionel Gulich and Yijie Guo and M. Gussert and Alex Hansen and Mihir Kulkarni and Chenran Li and Wei Liu and Viktor Makoviychuk and Grzegorz Malczyk and Hammad Mazhar and Masoud Moghani and Adithyavairavan Murali and Michael Noseworthy and Alexander Poddubny and Nathan Ratliff and Welf Rehberg and Clemens Schwarke and Ritvik Singh and James Latham Smith and Bingjie Tang and Ruchik Thaker and Matthew Trepte and Karl Van Wyk and Fangzhou Yu and Alex Millane and Vikram Ramasamy and Remo Steiner and Sangeeta Subramanian and Clemens Volk and CY Chen and Neel Jawale and Ashwin Varghese Kuruttukulam and Michael A. Lin and Ajay Mandlekar and Karsten Patzwaldt and John Welsh and Huihua Zhao and Fatima Anes and Jean-Francois Lafleche and Nicolas Moënne-Loccoz and Soowan Park and Rob Stepinski and Dirk Van Gelder and Chris Amevor and Jan Carius and Jumyung Chang and Anka He Chen and Pablo de Heras Ciechomski and Gilles Daviet and Mohammad Mohajerani and Julia von Muralt and Viktor Reutskyy and Michael Sauter and Simon Schirm and Eric L. Shi and Pierre Terdiman and Kenny Vilella and Tobias Widmer and Gordon Yeoman and Tiffany Chen and Sergey Grizan and Cathy Li and Lotus Li and Connor Smith and Rafael Wiltz and Kostas Alexis and Yan Chang and David Chu and Linxi \"Jim\" Fan and Farbod Farshidian and Ankur Handa and Spencer Huang and Marco Hutter and Yashraj Narang and Soha Pouya and Shiwei Sheng and Yuke Zhu and Miles Macklin and Adam Moravanszky and Philipp Reist and Yunrong Guo and David Hoeller and Gavriel State},\n  journal={arXiv preprint arXiv:2511.04831},\n  year={2025},\n  url={https:\u002F\u002Farxiv.org\u002Fabs\u002F2511.04831}\n}\n```\n\n## 致谢\n\nIsaac Lab 的开发始于 [Orbit](https:\u002F\u002Fisaac-orbit.github.io\u002F) 框架。我们衷心感谢 Orbit 的作者们所作出的基础性贡献。","# Isaac Lab 快速上手指南\n\nIsaac Lab 是一个基于 NVIDIA Isaac Sim 构建的 GPU 加速开源框架，旨在统一和简化机器人研究工作流（如强化学习、模仿学习和运动规划）。它结合了快速准确的物理引擎和传感器模拟，是机器人“仿真到现实”（Sim-to-Real）迁移的理想选择。\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**:\n    *   Linux (推荐 Ubuntu 22.04)\n    *   Windows 10\u002F11 (64-bit)\n*   **Python 版本**: 3.11\n*   **核心依赖**:\n    *   **NVIDIA Isaac Sim**: Isaac Lab 强依赖于特定版本的 Isaac Sim。请根据您的 Isaac Lab 版本选择对应的 Isaac Sim 版本：\n        *   `main` 分支 \u002F `v2.3.X`: 支持 Isaac Sim 4.5 \u002F 5.0 \u002F 5.1\n        *   `v2.2.X`: 支持 Isaac Sim 4.5 \u002F 5.0\n        *   `v2.0.X` - `v2.1.X`: 支持 Isaac Sim 4.5\n    *   **GPU**: 支持 CUDA 的 NVIDIA 显卡（建议 RTX 系列或更高），需安装正确的 NVIDIA 驱动程序。\n*   **其他工具**:\n    *   Git\n    *   Conda 或 Venv (用于管理 Python 虚拟环境)\n\n> **注意**: 中国大陆用户建议在下载 Isaac Sim 及相关大文件时，检查是否有国内镜像源或加速方案，以确保下载速度和稳定性。\n\n## 2. 安装步骤\n\n以下是基于本地安装的简要步骤。更详细的教程请参考 [官方文档](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fmain\u002Fsource\u002Fsetup\u002Finstallation\u002Findex.html#local-installation)。\n\n### 第一步：克隆仓库\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab.git\ncd IsaacLab\n```\n\n### 第二步：配置 Isaac Sim\n\n确保您已安装与当前分支兼容的 Isaac Sim 版本。如果您尚未安装，请访问 [NVIDIA Omniverse](https:\u002F\u002Fdocs.isaacsim.omniverse.nvidia.com\u002Flatest\u002Findex.html) 下载对应版本。\n\n设置环境变量以指向您的 Isaac Sim 安装路径（以 Linux 为例）：\n\n```bash\nexport ISAACSIM_PATH=${HOME}\u002F.local\u002Fshare\u002Fov\u002Fpkg\u002Fisaac-sim\n# 或者根据您的实际安装路径修改\n# export ISAACSIM_PATH=\u002Fpath\u002Fto\u002Fyour\u002Fisaac-sim\n```\n\n### 第三步：创建虚拟环境并安装依赖\n\n使用提供的脚本创建 conda 环境并安装 Isaac Lab 核心包：\n\n```bash\n# 创建名为 isaaclab 的 conda 环境\nconda create -n isaaclab python=3.11\nconda activate isaaclab\n\n# 安装 Isaac Lab 及其依赖\n# 注意：此步骤可能需要较长时间，因为它会编译部分组件\npython -m pip install -e .\n```\n\n### 第四步：验证安装\n\n运行一个简单的脚本来验证环境是否配置成功：\n\n```bash\n# 运行一个基础的空白环境测试\npython source\u002Fstandalone\u002Ftutorials\u002F00_sim\u002Fcreate_empty.py\n```\n\n如果成功启动 Isaac Sim 窗口且无报错，则安装完成。\n\n## 3. 基本使用\n\nIsaac Lab 内置了超过 30 个即开即用的训练环境，支持 RSL RL, SKRL, RL Games, Stable Baselines 等主流强化学习框架。\n\n### 运行预置的强化学习示例\n\n以下命令演示如何在一个经典的机械臂环境中训练一个强化学习策略（以 `FrankaCabinet` 环境为例，使用 RSL RL 框架）：\n\n```bash\n# 进入源码目录\ncd source\n\n# 运行训练脚本\n# --task: 指定任务名称\n# --headless: (可选) 无头模式运行，不显示图形界面，适合服务器训练\n# --num_envs: (可选) 并行环境数量，利用 GPU 加速\npython scripts\u002Freinforcement_learning\u002Frsl_rl\u002Ftrain.py --task Isaac-Franka-Cabinet-v0 --headless --num_envs 4096\n```\n\n### 自定义环境开发\n\n您可以基于现有的模板快速创建新环境。主要步骤包括：\n\n1.  在 `source\u002Fextensions\u002Fisaaclab_tasks\u002Fisaaclab_tasks\u002Fmanager_based\u002F` 下创建新的任务配置。\n2.  定义机器人的资产（URDF\u002FUSD）、观测空间（Observations）和动作空间（Actions）。\n3.  编写奖励函数（Reward Functions）。\n\n参考官方 [Tutorials](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fmain\u002Fsource\u002Ftutorials\u002Findex.html) 获取分步指南。\n\n---\n\n**更多资源**:\n*   [完整文档](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab)\n*   [可用环境列表](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fmain\u002Fsource\u002Foverview\u002Fenvironments.html)\n*   [技术论文 (arXiv)](https:\u002F\u002Farxiv.org\u002Fabs\u002F2511.04831)","某机器人初创团队正致力于训练一款四足机器人在复杂废墟环境中自主行走的强化学习策略。\n\n### 没有 IsaacLab 时\n- **仿真速度极慢**：在 CPU 上运行物理仿真，单次强化学习迭代需数小时，导致算法验证周期长达数周，严重拖慢研发进度。\n- **传感器模拟失真**：缺乏高精度的 RTX 光线追踪相机和激光雷达模拟，导致虚拟训练数据与真实世界差异巨大，模型难以迁移到真机（Sim-to-Real 失败）。\n- **环境搭建繁琐**：每次更换机器人构型或任务场景，都需要重写大量底层代码来配置物理属性和传感器接口，重复劳动占比过高。\n- **并发能力受限**：无法轻松并行运行数千个环境实例，难以收集足够的多样化数据来训练鲁棒的策略网络。\n\n### 使用 IsaacLab 后\n- **GPU 加速提效**：利用 GPU 并行计算能力，将仿真速度提升数十倍，原本需要一周的训练任务现在仅需几小时即可完成快速迭代。\n- **高保真感知模拟**：直接调用内置的 RTX 摄像头和激光雷达模块，生成逼真的深度图与点云数据，显著缩小了虚实差距，真机部署成功率大幅提升。\n- **统一框架复用**：基于预置的 16+ 种机器人模型和 30+ 个标准环境，团队只需关注核心算法逻辑，无需重复造轮子，新场景搭建时间从几天缩短至几小时。\n- **大规模并行训练**：轻松在本地或云端分布式启动上千个并行环境，高效采集海量交互数据，使策略能从容应对各种极端地形挑战。\n\nIsaacLab 通过 GPU 加速的高保真仿真与标准化工作流，将机器人强化学习的研发效率从“周级”提升至“小时级”，真正打通了从虚拟训练到现实落地的最后一公里。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fisaac-sim_IsaacLab_a7952be5.jpg","isaac-sim","NVIDIA Isaac Sim","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Fisaac-sim_4316e067.jpg","",null,"https:\u002F\u002Fgithub.com\u002Fisaac-sim",[83,87,90,94,98,102],{"name":84,"color":85,"percentage":86},"Python","#3572A5",98.2,{"name":88,"color":80,"percentage":89},"Kit",0.9,{"name":91,"color":92,"percentage":93},"Shell","#89e051",0.5,{"name":95,"color":96,"percentage":97},"Batchfile","#C1F12E",0.3,{"name":99,"color":100,"percentage":101},"Jinja","#a52a22",0.1,{"name":103,"color":104,"percentage":105},"Dockerfile","#384d54",0,6847,3341,"2026-04-06T15:44:40","BSD-3-Clause",4,"Linux, Windows","必需 NVIDIA GPU (基于 Isaac Sim)，具体型号和显存未说明，需支持 RTX 渲染","未说明",{"notes":115,"python":116,"dependencies":117},"该工具基于 NVIDIA Isaac Sim 构建，Isaac Sim 包含专有许可组件。不同版本的 Isaac Lab 对应不同的 Isaac Sim 版本（例如 main 分支支持 Isaac Sim 4.5\u002F5.0\u002F5.1）。macOS 未在支持平台列表中。若使用 isaaclab_mimic 扩展，需额外注意 cuRobo 的专有许可条款。","3.11",[118,119,120,121,122,123],"NVIDIA Isaac Sim (4.5 \u002F 5.0 \u002F 5.1)","RSL RL","SKRL","RL Games","Stable Baselines","cuRobo (仅 isaaclab_mimic 扩展需要)",[28,13],[126,127,128,76],"robot-learning","robotics","omniverse-kit-extension","2026-03-27T02:49:30.150509","2026-04-07T06:13:36.264555",[132,137,142,147,151,155],{"id":133,"question_zh":134,"answer_zh":135,"source_url":136},21032,"导入 URDF 文件失败或报错，可能是什么原因？","这通常是由网格文件（mesh files）的命名问题引起的。Isaac Sim 的 URDF 导入器不允许文件名中包含连字符（-）或其他特殊字符。例如，将 \"331839-DUMMY-M.stl\" 重命名为简单的名称如 \"bracket.stl\" 通常可以解决问题。请检查并简化所有引用网格文件的文件名。","https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fissues\u002F1760",{"id":138,"question_zh":139,"answer_zh":140,"source_url":141},21033,"运行代码时出现 'ModuleNotFoundError: No module named omni.isaac.kit' 错误怎么办？","这通常是因为 Python 环境变量未正确配置。你需要修改 `setup_python_env.sh` 文件（该文件通常被 `setup_conda_env_zsh.sh` 引用），确保 `PYTHONPATH` 包含了 Isaac Sim 的关键路径。具体需要添加的路径包括：`$SCRIPT_DIR\u002Fkit\u002Fpython\u002Flib\u002Fpython3.10\u002Fsite-packages`、`$SCRIPT_DIR\u002Fexts\u002Fomni.isaac.kit`、`$SCRIPT_DIR\u002Fkit\u002Fkernel\u002Fpy` 等。修改后重新激活虚拟环境或重新运行设置脚本。","https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fissues\u002F516",{"id":143,"question_zh":144,"answer_zh":145,"source_url":146},21034,"Observation Manager 是否支持观察历史（Observation Histories）功能？","目前原生的 Observation Manager 默认不支持在组级别（group level）直接存储历史数据，通常只支持按项（term level）堆叠（格式如 AABBCC）。如果需要实现历史功能，用户可以自定义一个包装类（Wrapper Class），继承 `ManagerTermBase`，在内部维护一个张量来存储历史帧，并在 `reset` 和调用时更新该张量。对于需要特定架构（如 CNN 或 Transformer）处理的历史堆叠顺序，需在自定义实现中注意数据维度排列。","https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fissues\u002F1208",{"id":148,"question_zh":149,"answer_zh":150,"source_url":136},21035,"Isaac Sim 新版本更新后，之前的 URDF 导入代码突然失效了，如何排查？","首先检查是否遇到了已知的命名限制问题（如文件名含连字符）。其次，确认 Isaac Sim 版本变更是否引入了新的 API 或配置要求。如果问题依旧存在且文件名已简化，建议查看 NVIDIA 开发者论坛关于特定版本（如 4.5.0 或 5.1）的 URDF 导入器变更日志。在某些情况下，即使重命名文件也可能因缓存或特定版本 Bug 导致问题持续，尝试清理缓存或检查是否有针对该版本的补丁。",{"id":152,"question_zh":153,"answer_zh":154,"source_url":146},21036,"如何在 IsaacLab 中手动实现观察历史（History）以用于 RMA 或 Student-Teacher 方法？","你可以创建一个自定义类继承自 `ManagerTermBase`。在 `__init__` 中初始化一个形状为 `(num_envs, history_len, obs_dim)` 的零张量用于存储历史数据。在 `reset` 方法中，用当前观测值填充该张量。在每次步长中，将新观测值推入历史队列并移除最旧的值（或使用滚动更新）。最后，将处理后的历史张量返回作为观测值。注意这种方法是在每个观测项（term）级别实现的，如果需要组级别堆叠，需调整管理器逻辑。",{"id":156,"question_zh":157,"answer_zh":158,"source_url":141},21037,"解决 'omni.isaac.kit' 模块找不到错误的具体环境变量配置是什么？","需要在 `setup_python_env.sh` 中导出以下 `PYTHONPATH`（根据实际安装路径调整 `$SCRIPT_DIR`）：\n`export PYTHONPATH=$PYTHONPATH:$SCRIPT_DIR\u002F..\u002F..\u002F..\u002F$PYTHONPATH:$SCRIPT_DIR\u002Fkit\u002Fpython\u002Flib\u002Fpython3.10\u002Fsite-packages:$SCRIPT_DIR\u002Fpython_packages:$SCRIPT_DIR\u002Fexts\u002Fomni.isaac.kit:$SCRIPT_DIR\u002Fkit\u002Fkernel\u002Fpy:$SCRIPT_DIR\u002Fkit\u002Fplugins\u002Fbindings-python:$SCRIPT_DIR\u002Fexts\u002Fomni.isaac.lula\u002Fpip_prebundle:$SCRIPT_DIR\u002Fexts\u002Fomni.exporter.urdf\u002Fpip_prebundle:$SCRIPT_DIR\u002Fextscache\u002Fomni.kit.pip_archive-*\u002Fpip_prebundle:$SCRIPT_DIR\u002Fexts\u002Fomni.isaac.core_archive\u002Fpip_prebundle:$SCRIPT_DIR\u002Fexts\u002Fomni.isaac.ml_archive\u002Fpip_prebundle`。\n同时确保 `LD_LIBRARY_PATH` 也正确设置。",[160,165,170,175,180,185,190,195,200,205,210,215,220,225,230,235,240,245,250,255],{"id":161,"version":162,"summary_zh":163,"released_at":164},127102,"v3.0.0-beta","# Isaac Lab 3.0 Beta 🚀\n\nIsaac Lab 3.0 Beta 是 Isaac Lab 的下一个重大版本，基于 **Isaac Sim 6.0** 构建，并进行了从头开始的架构重构。此版本引入了多后端物理引擎、可插拔渲染器系统、Warp 原生数据流水线以及无工具包安装模式——从而实现更快、更灵活的机器人学习研究。\n\n> ⚠️ **这是一个测试版。** `develop` 分支目前仍在积极开发中，在某些用例中可能会出现破坏性变更、错误信息或性能退化。\n\n---\n\n## ✨ 亮点\n\n### 多后端物理引擎架构\nIsaac Lab 3.0 引入了基于工厂模式的多后端架构，将与仿真后端相关的代码与核心 API 彻底分离。资产和传感器类（例如 `Articulation`、`RigidObject`、`ContactSensor`）现在由抽象基类提供支持，具体的后端实现则位于专用的扩展包中：\n\n- **`isaaclab_physx`** — 完整的 PhysX 后端（默认），支持可变形物体、表面夹爪、接触传感器、IMU 和帧变换器。\n- **`isaaclab_newton`** — 基于 MuJoCo-Warp 的全新 Newton 物理后端，支持 MJWarp、XPBD 和 Featherstone 求解器，并通过 CUDA 图加速。\n\n您现有的来自 `isaaclab.assets` 和 `isaaclab.sensors` 的导入将继续正常工作——工厂会在运行时自动调度到当前激活的后端。\n\n### Newton 物理后端\n新的 `isaaclab_newton` 扩展允许在 **无需 Isaac Sim** 的情况下运行 Isaac Lab 环境（无工具包模式）。Newton 支持包括：\n\n- 关节体、刚体及刚体集合\n- 接触传感器\n- MuJoCo-Warp 求解器，支持可配置的积分器（`implicitfast`、`euler`）和接触模型（`pyramidal`、`elliptic`）\n- CUDA 图支持，用于高吞吐量的步进计算\n- 适用于 20 多种环境的 Newton 兼容预设（如运动、操作和经典控制任务）\n\n### 可插拔渲染器系统\n一个新的 `BaseRenderer` 抽象层通过工厂模式支持多种渲染后端：\n\n| 后端         | 是否需要 Isaac Sim？ | 适用场景                     |\n|--------------|----------------------|------------------------------|\n| **Isaac RTX** | 是                   | 高保真度传感器、照片级真实感渲染 |\n| **OVRTX**     | 否（需 `isaaclab_ov[ovrtx]`） | 无工具包 RTX 流水线           |\n| **Newton Warp** | 否（无工具包）       | 无需 RTX 的快速训练            |\n\n### 可插拔可视化系统\n\nIsaac Lab 3.0 引入了一个全新的 **可插拔可视化框架**（`isaaclab_visualizers`），包含四个可互换的后端，均与物理引擎和渲染器解耦：\n\n| 可视化工具   | 适用场景                 | 主要特性                       |\n|--------------|--------------------------|--------------------------------|\n| **Omniverse (Kit)** | 高保真度、与 Isaac Sim 集成 | USD 场景、视觉标记、实时训练图表 |\n| **Newton**    | 快速迭代、低开销         | OpenGL 渲染、物理调试标记（关节、接触、质心）、交互式相机控制 |\n| **Rerun**     | 远程查看与回放          | Web 查看器、时间轴滑动、`.rrd` 格式记录\u002F导出 |\n| **Vis","2026-03-17T06:38:37",{"id":166,"version":167,"summary_zh":168,"released_at":169},127103,"v2.3.2","## 变更内容\n\n本次发布主要聚焦于稳定性、基础设施改进、工作流优化以及功能的逐步扩展，同时还引入了一些重要的新特性，包括 **无人机多旋翼和推进器支持**、**多网格 RayCaster**、**基于视觉的触觉传感器**、**Haply 设备集成** 以及全新的 **OpenArm 环境**。\n\n此外，本次版本还对训练工作流、遥操作与 Mimic 流水线、Ray 集成、仿真工具以及开发者工具链进行了优化，并修复了大量提升鲁棒性和改善开发体验的问题。\n\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F19624490-b9ef-41d4-8a74-67ccf96fdaed\n\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F3222f88d-46ee-4816-8d73-d8910b83d4a8\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fcompare\u002Fv2.3.1...v2.3.2\n\n> [!NOTE]\n> 这将是当前 main 分支的最后一个版本，因为我们已将开发重心转移到 develop 分支。\n>\n> 在迈向 Isaac Lab 3.0 的过程中，我们计划对 [`develop`](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Ftree\u002Fdevelop) 分支进行重大重构。我们欢迎社区继续贡献，但未来的活跃开发将主要在 develop 分支上进行。\n>\n> 如果您有未合并的 PR，请将其重新指向 develop 分支，以确保其与最新更改保持一致。\n\n## ✨ 新特性\n\n### 核心与仿真\n\n* @renezurbruegg 在 #3298 中添加了支持动态网格追踪的 RayCaster\n* @JuanaDd 在 #3420 中添加了基于视觉的触觉传感器及形状感知示例\n* @AntoineRichard 在 #3287 中添加了力矩合成器，允许在同一刚体上组合多个力矩\n* @mihirk284、@grzemal 和 @Zwoelf12 在 #3760 中添加了多旋翼\u002F推进器执行器、多旋翼资产以及基于管理器的 ARL 无人机任务\n* @bmccann-bdai 在 #3864 中添加了 IMU 传感器的自动变换发现功能，用于查找有效的父级刚体\n* @gattra-rai 在 #3563 中为 ContactSensor 添加了摩擦力报告功能\n* @KyleM73 在 #1672 中添加了用于导入 MJCF 基础资产的 MJCF 生成器\n\n### 学习与环境\n\n* @JinnnK 在 #4089 中添加了 OpenArm 环境\n\n### Mimic 与遥操作\n\n* @mingxueg-nv 在 #3873 中添加了带有力反馈和遥操作演示的 Haply 设备 API\n* @rwiltz 在 #3950 中重构了重定向器，并为 G1 任务添加了 Quest 重定向器\n* @rwiltz 在 #4140 中添加了 Arena G1 位姿操作重定向器\n* @peterd-NV 在 #3992 中为 Isaac Lab Mimic 添加了用于生成位姿操作数据的 API\n\n## 🔧 改进\n\n### 核心与仿真\n\n* @renezurbruegg 在 #3716 中为 JointPositionToLimitsAction 添加了 preserve-order 标志\n* @Mayankm96 在 #3367 中为图元获取工具添加了实例化网格的解析功能\n* @kellyguo11 在 #3391 中为环境添加了可配置的日志目录参数\n* @ooctipus 在 #3502 中通过 PhysxCfg 暴露了 PhysX 标志 solveArticulationContactLast\n* @kellyguo11 在 #3709 中移除了配置加载和保存中的 pickle 依赖\n* 改进了 ","2026-02-02T18:54:35",{"id":171,"version":172,"summary_zh":173,"released_at":174},127104,"v2.3.1","## 变更内容\n\n这是一个小型补丁版本，包含几项影响用户工作流的重要修复。\n\n主要修复包括：\n* 在基于管理器的工作流中，终止日志记录的行为已更改：`get_done_term` 现在返回当前步骤的值，而不是上一集的值。\n* 此外，Isaac Sim 5.1 引入了 URDF 导入器的一项破坏性变更，即不再支持合并关节标志。我们现已在导入器中添加补丁以恢复原有行为。未来，我们将弃用该标志，转而直接保留来自 URDF 的资产定义，无需在导入过程中进行额外处理。\n\n## 🐛 错误修复\n\n* 将 URDF 导入器更新至 2.4.31，以继续支持合并关节功能，由 @kellyguo11 在 #4000 中实现\n* 将每步终止与最后一集终止的记账逻辑分离，由 @ooctipus 在 #3745 中实现\n* 如果执行器配置中未指定力限制，则使用 USD 中的 `effort_limit` 值，由 @JuanaDd 在 #3522 中实现\n* 修复 `from_files` 配置中肌腱属性的类型名称错误，由 @KyleM73 在 #3941 中实现\n* 修复 pip 安装文档中的重复文本，由 @shryt 在 #3969 中实现\n* 锁定 `pre-commit.yaml` 工作流的 Python 版本，由 @hhansen-bdai 在 #3929 中实现\n\n## 📜 文档\n\n* 更新模仿遥控操作文档，使其链接到运动策略训练，由 @huihuaNvidia2023 在 #4053 中实现\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fcompare\u002Fv2.3.0...v2.3.1","2025-12-04T22:26:35",{"id":176,"version":177,"summary_zh":178,"released_at":179},127105,"v2.3.0","## 变更内容\n\nIsaac Lab 2.3.0 版本基于 Isaac Sim 5.1 构建，针对灵巧操作、遥操作和学习工作流进行了全面增强。该版本引入了具备先进训练能力的新灵巧环境，扩展了表面夹爪和遥操作对更多机器人及设备的支持，并将 SkillGen 与 Mimic 仿生学习流水线集成，通过 cuRobo 集成实现 GPU 加速的运动规划和基于技能的数据生成。\n\n本次发布的主要亮点包括：\n\n- **灵巧强化学习（DexSuite）**：推出了两个基于 Kuka 机械臂和 Allegro 手部配置的新灵巧操作环境，并新增对自动领域随机化（ADR）和基于种群的训练（PBT）的支持。\n- **表面夹爪更新**：表面夹爪现已扩展支持基于 Manager 的工作流，新增了 `SurfaceGripperAction` 和 `SurfaceGripperActionCfg`，并提供了多个新环境，演示使用表面夹爪和 RMPFlow 控制器进行遥操作的示例。同时引入了多款新机器人及其变体，包括配备 Robotiq 夹爪和吸盘的 Franka 和 UR10，以及 Galbot 和 Agibot 机器人。\n- **Mimic - SkillGen**：为 Mimic 仿生学习流水线增加了 SkillGen 支持，引入 cuRobo 集成，将 GPU 运动规划与基于技能片段的数据生成相结合。请注意，cuRobo 采用专有许可条款，请在使用前仔细阅读 [cuRobo 许可协议](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fblob\u002Fmain\u002Fdocs\u002Flicenses\u002Fdependencies\u002FcuRobo-license.txt)。\n- **Mimic - 定位操纵**：新增了一个 G1 类人型环境，将基于强化学习的行走与基于逆运动学的操作相结合。该环境集成了完整的机器人导航栈，可通过随机化桌面上的抓取\u002F放置位置、目标点以及地面障碍物来丰富演示数据。通过将任务划分为抓取-导航-放置三个阶段，此方法能够从仅包含操作动作的演示中生成大规模的定位操纵数据集。\n- **遥操作**：改进了上半身逆运动学控制器，新增了一个零空间姿态任务，可在执行类人型任务时支持腰部运动，同时将冗余自由度规整为首选的直立姿势。此外，还新增了对 Vive 和 Manus 手套的支持，为遥操作设备提供了更多选择。\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fcompare\u002Fv2.2.1...v2.3.0\n\n## 🔆 重点功能\n\n### Dex 环境\n\n我们新增了一个 DexSuite，其中包括遵循 [DextrAH](https:\u002F\u002Fdextrah-rgb.github.io\u002F) 和 [DexPBT](https:\u002F\u002Farxiv.org\u002Fabs\u002F2305.12127) 标准的灵巧提升和重定向环境。这些环境还展示了自动领域随机化（ADR）和基于种群的训练（PBT）的使用方法。\n\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F89cce6ef-a029-4705-8ef7-9ad5efee6320\n\n\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments","2025-10-28T21:38:54",{"id":181,"version":182,"summary_zh":183,"released_at":184},127106,"v2.2.1","## 👀 概述\n\n这是一个包含一些改进和错误修复的次要补丁版本。\n\n**完整更新日志**: https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fcompare\u002Fv2.2.0...v2.2.1\n\n## ✨ 新特性\n\n* 由 @jtigue-bdai 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2842 中为 ContactSensor 添加了接触点位置报告功能。\n* 由 @AntoineRichard 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2730 中添加了用于导出的环境动作\u002F观测描述符。\n* 由 @Mayankm96 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3057 中添加了 RSL-RL 对称性示例，适用于 cartpole 和 ANYmal 行走任务。\n\n## 🔧 改进\n\n### 核心 API\n\n* 由 @michaellin6 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3149 中增强了 Pink IK 控制器，增加了零空间姿态控制等功能。\n* 由 @matthewtrepte 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3221 中添加了在 Nucleus 服务器上检查 USD 路径时的周期性日志记录。\n* 由 @ooctipus 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3110 中禁止在 process_sb3_cfg 中评估 sb3_ppo_cfg.yaml 文件中写入的字符串值。\n\n### 基础设施\n\n* **应用设置**\n  * 由 @matthewtrepte 和 @kellyguo11 分别在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3219 和 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3089 中禁用了无头模式及无头渲染应用的速率限制。\n  * 由 @matthewtrepte 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3240 中禁用了渲染预设“平衡”和“性能”模式下的 `rtx.indirrectDiffuse.enabled` 设置。\n  * 由 @soowanpNV 和 @rwiltz 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3238 和 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3255 中将性能分析器后端默认设置为 NVTX。\n* **依赖项**\n  * 由 @hhansen-bdai 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3116 中添加了 hf-xet 许可证。\n  * 由 @kellyguo11 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3237 中修复了新的 typing-inspection 依赖项许可证问题。\n* **测试与基准测试**\n  * 由 @louislelay 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3058 中添加了基于缩放随机化的范围的基本验证测试。\n  * 由 @jtigue-bdai 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3160 中添加了 `SensorBase` 测试。\n* **仓库工具**\n  * 由 @hhansen-bdai 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3104 中改进了 install_deps.py 的输出读取功能。\n  * 由 @ooctipus 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3139 中修复了 isaaclab.sh，使其能够准确检测 isaacsim_version 是否为 4.5 或 ≥ 5.0。\n  * 由 @ooctipus 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3291 中禁用了 conftest.py 中的详细打印输出。\n  * 由 @ben-johnston-nv 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3247 中更新了 pytest 标志，以支持 isaacsim 集成测试。\n  * 由 @pascal-roth 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3090 中细化了 CodeOwners 文件。\n  * 由 @nv-apoddubny 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F3120 中修复了 CI 中的一些小问题。\n\n## 🐛 错误修复\n\n### 核心 API\n\n* **资产接口**\n  * 修复了在关节处设置摩擦系数到 PhysX 中的问题。","2025-08-29T17:58:55",{"id":186,"version":187,"summary_zh":188,"released_at":189},127107,"v2.2.0","## 👀 概述\n\n**Isaac Lab 2.2** 在仿真能力、工具链和开发者体验方面带来了重大升级。它扩展了对高级物理特性、新环境以及改进的测试和文档工作流的支持。此版本与 **Isaac Sim 5.0** 完全兼容，同时也向后兼容 **Isaac Sim 4.5**。\n\n本次发布的主要亮点包括：\n\n- **增强的物理支持**：使用最新的 PhysX API 更新了 [关节摩擦建模](https:\u002F\u002Fnvidia-omniverse.github.io\u002FPhysX\u002Fphysx\u002F5.6.1\u002Fdocs\u002FArticulations.html#articulation-joint-friction)，新增了对 [空间肌腱](https:\u002F\u002Fnvidia-omniverse.github.io\u002FPhysX\u002Fphysx\u002F5.6.1\u002Fdocs\u002FArticulations.html#spatial-tendons) 的支持，并改进了表面抓取器的交互。\n- **用于模仿学习的新环境**：引入了两个新的 GR1 模仿环境，具备领域随机化和视觉鲁棒性评估功能，并优化了拾取与放置任务。\n- **新的接触密集型操作任务**：集成了 [FORGE](https:\u002F\u002Fnoseworm.github.io\u002Fforge\u002F) 和 [AutoMate](https:\u002F\u002Fbingjietang718.github.io\u002Fautomate\u002F) 任务，用于在仿真中学习细粒度的接触交互。\n- **远程操作改进**：远程操作工具经过增强，提供了可配置参数和 CloudXR 运行时更新，包括头部跟踪和手部跟踪功能。\n- **性能与易用性提升**：支持内存中的 Stage 和 Fabric 中的克隆功能，以加快场景创建；新增 OVD 录制器，用于大场景的基于 GPU 的动画录制；并引入 FSD（Fabric 场景代理），以提升渲染速度。\n- **文档改进**：文档得到了扩展和更新，涵盖了新功能、常见问题的解决方法以及流程简化，包括远程操作系统要求、VS Code 集成和 Python 环境管理等方面的更新。\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fcompare\u002Fv2.1.1...v2.2.0\n\n## 🔆 重点功能\n\n### [FORGE 环境](https:\u002F\u002Fnoseworm.github.io\u002Fforge\u002F)\n\n如 [FORGE](https:\u002F\u002Fnoseworm.github.io\u002Fforge\u002F) 论文中所介绍，这些环境专注于需要力感知操作的多阶段装配任务。更多详情请参阅 PR #2968。\n\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F0f307974-d0e8-40c4-a5b0-2f5c5d292668\n\n### [AutoMate 环境](https:\u002F\u002Fbingjietang718.github.io\u002Fautomate\u002F)\n\n[Automate](https:\u002F\u002Fbingjietang718.github.io\u002Fautomate\u002F) 项目旨在创建一个大型装配数据集，用于训练覆盖广泛装配任务的策略。目前，所有 100 个装配部件及其对应的环境均已纳入环境套件。更多信息请参阅 PR #2507。\n\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F50bd0d4c-e281-405f-bdb2-6b77854ed0bb\n\n### [Isaac Lab 评估任务](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLabEvalTasks)\n\n用于模拟和评估操作策略（例如 [Isaac GR00T N1](https:\u002F\u002Fgit","2025-08-07T19:42:33",{"id":191,"version":192,"summary_zh":193,"released_at":194},127108,"v2.1.1","## 👀 概述\n\n本次发布在过去几个月中持续开发，涵盖了整个代码库中的大量更新、增强和新功能。鉴于变更数量较多，我们将其按相关类别进行了分组，以提升可读性。该版本与 [NVIDIA Isaac Sim 4.5](https:\u002F\u002Fdocs.isaacsim.omniverse.nvidia.com\u002F4.5.0\u002Finstallation\u002Fdownload.html) 兼容。\n\n我们感谢社区在整个过程中展现出的耐心与贡献，这些都为确保质量和稳定性发挥了重要作用。未来我们将致力于更频繁地发布补丁版本，以进一步提升开发者体验。\n\n**注意：** 此次小版本不包含 Docker 镜像或 pip 包。\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fcompare\u002Fv2.1.0...v2.1.1\n\n## ✨ 新特性\n\n* **资产接口**\n  * 添加了 `position` 参数，用于在刚体的不同位置施加外力和扭矩，由 @AntoineRichard 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1680 中实现。\n  * 向 ArticulationData 字段添加了 `body_incoming_joint_wrench_b`，由 @jtigue-bdai 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2128 中实现。\n  * 允许显式选择关节树根 Prim，由 @lgulich 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2228 中实现。\n* **传感器接口**\n  * 为 FrameTransformer 可视化添加连接线，由 @Mayankm96 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1754 中实现。\n  * 在 FrameTransformer 内部使用可视化标记来绘制连接线，由 @bikcrum 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2526 中实现。\n* **MDP 项**\n  * 添加了 `body_pose_w` 和 `body_projected_gravity_b` 观测值，由 @jtigue-bdai 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2212 中实现。\n  * 添加了关节力矩观测值，由 @jtigue-bdai 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2211 中实现。\n  * 将 CoM 随机化项添加到基于管理器的事件中，由 @shendredm 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1714 中实现。\n  * 添加了基于时间的 MDP（观测）函数，由 @TheIndoorDad 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2332 中实现。\n  * 添加了课程学习 MDP 项，用于修改任何环境参数，由 @ooctipus 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2777 中实现。\n* **新增示例任务**\n  * 添加了来自 Automate 项目的装配任务，由 @yijieg 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2507 中实现。\n  * 添加了 Digit 行走示例，由 @lgulich 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1892 中实现。\n\n## 🔧 改进\n\n### 核心 API\n\n* **执行器接口**\n  * 修复了资产的隐式执行器限制配置，由 @ooctipus 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2952 中实现。\n  * 更新了 Franka 机械臂的执行器配置，由 @reeceomahoney 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2492 中实现。\n* **资产接口**\n  * 优化了资产类内部数据的 getter 方法，由 @Mayankm96 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2118 中实现。\n  * 添加了设置资产 Prim 可见性的方法，由 @Mayankm96 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1752 中实现。\n* **传感器接口*","2025-07-30T12:59:44",{"id":196,"version":197,"summary_zh":198,"released_at":199},127109,"v2.1.0","## 👀 概述\n\n本次发布正式支持使用 Apple Vision Pro 进行遥操作，以收集高质量且灵巧的手部数据，其中包括通过 Isaac Lab Mimic 实现的双手遥操作和模仿学习工作流。\n\n我们还引入了针对 USD 属性的全新随机化方法，涵盖缩放、颜色和纹理的随机化。在本版本中，RSL RL 更新至 v2.3.1，新增了多项功能，包括分布式训练、师生蒸馏以及循环师生蒸馏。\n\n此外，我们对 [Extension Template](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLabExtensionTemplate) 进行了全面重构，新增了一个可直接在 Isaac Lab 仓库内使用的自动模板生成工具。扩展模板为用户在自托管仓库中开发新项目提供了一种强大而灵活的方式，能够与核心 Isaac Lab 仓库及其变更保持隔离。此前的 IsaacLabExtensionTemplate 仓库仅提供了基于 Manager 工作流和 RSL RL 的有限示例；而在新的模板生成器中，用户可以从所有受支持的工作流和强化学习库中进行选择，并指定所需的强化学习算法。我们将在不久的将来弃用独立的 [IsaacLabExtensionTemplate](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLabExtensionTemplate) 仓库。\n\nNVIDIA 同时发布了 [HOVER](https:\u002F\u002Fgithub.com\u002FNVlabs\u002FHOVER)，作为一个独立的代码仓库，其中包含基于 Isaac Lab 构建的人形机器人全身神经控制器。HOVER 提供了从仿真到现实的部署流程，可用于 Unitree H1 机器人；我们也在 Isaac Lab 文档中新增了该部署过程的教程指南。\n\n**完整更新日志**: https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fcompare\u002Fv2.0.2...v2.1.0 \n\n## 🔆 亮点功能\n![gr-1_pick_place](https:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F332c203a-7ba7-4eb0-8df8-affe506f2091)\n\n[TextureRandomization.webm](https:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002Fb2e2f081-b194-44cf-b49e-0a961593e127)\n\n## ✨ 新特性\n* @Toni-SM 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2039 中新增了外部项目\u002F内部任务模板生成器。\n* @louislelay 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2221 中为外部任务模板生成器添加了虚拟智能体。\n* @Mayankm96 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2040 中为事件管理器新增了 USD 级别的随机化模式。\n* @hapatel-bdai 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2121 中新增了纹理和缩放随机化的事件项。\n* @Mayankm96 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2153 中新增了用于随机化颜色的 Replicator 事件。\n* @kellyguo11 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2041 中新增了 H1 行走的交互式演示脚本。\n* @chengronglai 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1944 中新增了 Franka 堆叠模仿的蓝图环境。\n* @Mayankm96 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F2019 中为 RSL RL 封装添加了动作裁剪功能。","2025-04-24T15:13:49",{"id":201,"version":202,"summary_zh":203,"released_at":204},127110,"v2.0.2","## 👀 概述\n\n此补丁版本专注于改进执行器配置并修复关键错误，同时回滚 v2.0.1 中未预期的行为变更。如果您正从 Isaac Lab 的 2.0 之前的版本迁移，**我们强烈建议切换**到这个新版本。\n\n**主要变更：**\n\n* **执行器限制处理**：引入了 `velocity_limit_sim` 和 `effort_limit_sim`，以明确区分仿真求解器的限制与执行器模型的约束。已将隐式执行器速度限制恢复为 v2.0 之前的默认行为。\n* **仿真配置更新**：移除了 `disable_contact_processing` 标志，以简化行为。\n* **渲染配置更新**：恢复为 v2.0 之前的配置，以提升渲染效果的质量。\n* **Tiled Camera 修复**：修复了运动向量处理，并添加了一个用于从 TiledCamera 获取语义图像的热修复补丁。\n* **WebRTC 支持**：增加了直播流的 IP 地址指定功能。\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fcompare\u002Fv2.0.1...v2.0.2\n\n## ✨ 新特性\n\n* 由 @jtigue-bdai 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1654 中为执行器添加了 `velocity_limit_sim` 和 `effort_limit_sim`。\n* 由 @sheikh-nv 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1967 中添加了支持 IP 地址指定的 WebRTC 直播流功能。\n\n## 🔧 改进\n\n* 由 @Mayankm96 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1876 中添加了代码贡献指南和示例。\n* 由 @Mayankm96 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1751 中将关节状态设置器分离到 Articulation 类内部。\n* 由 @Toni-SM 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1972 中实现了 skrl 多智能体算法的确定性评估。\n* 由 @Mayankm96 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1988 中为 `pyproject.toml` 添加了新的扩展。\n* 由 @j3soon 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1970 中更新了关于 Isaac Sim 二进制安装路径和 VSCode 集成的文档。\n* 由 @jtigue-bdai 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1851 中移除了 RigidObject 弃用中的剩余弃用警告。\n* 由 @kellyguo11 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1830 中向文档添加了安全性和说明性注释。\n* 由 @kellyguo11 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1943 中更新了关于分割和 50 系列 GPU 问题的文档。\n* 由 @kellyguo11 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1947 中为 tiled camera 的语义分割问题添加了临时解决方案。\n\n## 🐛 错误修复\n\n* 由 @peterd-NV 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1839 中修复了在使用并行环境时 Franka 堆叠环境中对象观测的偏移问题。\n* 由 @jtigue-bdai 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1809 和 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1865 中为 ManagerBasedEnv、DirectRLEnv 和 MARL 环境的初始化添加了场景更新。\n* 由 @Mayankm96 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1862 中以 eval() 模式加载执行器网络，以防止梯度计算。\n* 修复了导入 ANYm 的说明。","2025-03-05T23:37:20",{"id":206,"version":207,"summary_zh":208,"released_at":209},127111,"v2.0.1","## 👀 概述\n\n本次发布包含少量修复和改进。\n\n主要变更在于保持与 RSL RL 更新后库名的兼容性。此前 Isaac Lab 的安装方式因库名变更而失效，而本版本在 Isaac Lab 中提供了必要的修复和更新，以适应库名变化并确保与 RSL RL 安装的兼容性。\n\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fcompare\u002Fv2.0.0...v2.0.1\n\n## 🔧 改进\n* @Mayankm96 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1811 中将 RSL-RL 的安装方式切换为通过 PyPI 进行。\n* @fan-ziqi 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1766 中更新了文档中的脚本路径。\n* @kellyguo11 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1788 中禁用了保存文件时的扩展自动重载功能。\n* @kellyguo11 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1818 中更新了 v2.0.1 版本的安装文档。\n\n## 🐛 问题修复\n* @Jackkert 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1765 中修复了将关节运动学姿态写入仿真时，通信和链接缓冲区的时间戳问题。\n* @louislelay 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1776 中修复了 xdg-open 命令中本地文档预览路径错误的问题。\n* @samibouziri 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1808 中修复了找不到 rsl-rl 分发包（不可用）的问题。\n* @zoctipus 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1821 中修复了 RayCaster 传感器内部传感器漂移重置的问题。\n\n## 新贡献者\n* @Jackkert 在 https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1765 中完成了首次贡献。","2025-02-26T01:27:29",{"id":211,"version":212,"summary_zh":213,"released_at":214},127112,"v2.0.0","## 👀 Overview\r\n\r\nIsaac Lab 2.0 brings some exciting new features, including a new addition to the Imitation Learning workflow with the **Isaac Lab Mimic** extension.\r\n\r\nIsaac Lab Mimic provides the ability to automatically generate additional trajectories based on just a few human collected demonstrations, allowing for larger training datasets with less human effort. This work is based on the [MimicGen](https:\u002F\u002Fmimicgen.github.io\u002F) work for Scalable Robot Learning using Human Demonstrations.\r\n\r\nAdditionally, we introduced a new set of AMP tasks based on [Adversarial Motion Priors](https:\u002F\u002Fxbpeng.github.io\u002Fprojects\u002FAMP\u002Findex.html), training humanoid robots to walk, run, and dance 👯 \r\n\r\nAlong with Isaac Lab 2.0, Isaac Sim 4.5 brings several new and breaking changes, including a full refactor of the Isaac Sim extensions, an improved URDF importer, an update to the PyTorch dependency to version 2.5.1, and many fixes for tiled rendering that now supports multiple tiled cameras at different resolutions.\r\n\r\nTo follow the refactoring in Isaac Sim, we made similar refactoring and restructuring changes to Isaac Lab. These breaking changes will no longer be compatible with previous Isaac Sim versions. **Please make sure to update to Isaac Sim 4.5 when using the Isaac Lab 2.0 release**.\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fcompare\u002Fv1.4.1...v2.0.0\r\n\r\n## 🌟 Highlights from the Isaac Sim 4.5 release:\r\n\r\n- Support for multiple `TiledCamera` instances and varying resolutions\r\n- Improved rendering performance by up to 1.2x\r\n- Faster startup time through optimizations in the Cloner class that improves startup time by 30%\r\n- Enhanced OmniPVD for debugging physics simulation, enabling capturing reinforcement learning simulation workloads of up to 2000 environments\r\n- Physics simulation performance optimizations improving throughput of up to 70%\r\n- Physics support for dedicated cylinder and cone geometry designed for robot wheels that is fully GPU accelerated \r\n- A new physics GPU filtering mechanism allowing co-location of reinforcement learning environments at the origin with minimal performance loss for scenes with limited collider counts\r\n- Improvements in simulation stability for mimic joints at high joint gains\r\n\r\n## 🔆 Highlighted Features\r\n\r\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002Ffee52af0-2ef5-4753-8fbb-dd887f691967\r\n\r\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002Fac1310d4-5452-4bb0-a9d5-d4b8202d79d3\r\n\r\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002Fea1e7901-5711-4324-ae91-3649f6c4fcbd\r\n\r\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F288ffb91-20b7-4396-a1e1-1ee8f222d444\r\n\r\n\r\n## ✨ New Features\r\n\r\n* Adds humanoid AMP tasks for direct workflow by @Toni-SM \r\n* Adds Isaac Lab Mimic based on MimicGen data generation for Imitation Learning by @peterd-NV @nvcyc @ashwinvkNV @karsten-nvidia \r\n* Adds consolidated demo script for showcasing recording and mimic dataset generation in real-time in one simulation script by @nvcyc \r\n* Adds Franka stacking environment for GR00T mimic by @peterd-NV @nvcyc \r\n* Adds option to filter collisions and real-time playback by @kellyguo11 \r\n\r\n## 🔧 Improvements\r\n\r\n* Adds a tutorial for policy inference in a prebuilt USD scene by @oahmednv\r\n* Adds unit tests for multi-tiled cameras by @matthewtrepte  \r\n* Updates render setting defaults for better quality by @kellyguo11 \r\n* Adds a flag to wait for texture loading completion when reset by @oahmednv \r\n* Adds pre-trained checkpoints and tools for generating and uploading checkpoints by @nv-cupright \r\n* Adds new denoiser optimization flags for rendering by @kellyguo11 \r\n* Updates torch to 2.5.1 by @kellyguo11 \r\n* Renames default conda and venv Python environment from ``isaaclab`` to ``env_isaaclab`` by @Toni-SM\r\n* \r\n## 🐛 Bug Fixes\r\n\r\n* Fixes external force buffers to set to zero when no forces\u002Ftorques are applied by @matthewtrepte \r\n* Fixes RSL-RL package name in ``setup.py`` according to PyPI installation by @samibouziri\r\n\r\n## 💔 Breaking Changes\r\n\r\n* Updates the URDF and MJCF importers for Isaac Sim 4.5 by @Dhoeller19 \r\n* Renames Isaac Lab extensions and folders by @kellyguo11 \r\n* Restructures extension folders and removes old imitation learning scripts by @kellyguo11\r\n* Renames default conda and venv Python environment from `isaaclab` to `env_isaaclab` by @Toni-SM\r\n\r\n## ✈️ Migration Guide\r\n\r\n### Renaming of Isaac Sim Extensions\r\n\r\n\u003Cdetails>\r\n\u003Csummary>Details\u003C\u002Fsummary>\r\n\r\nPreviously, Isaac Sim extensions have been following the convention of `omni.isaac.*`,\r\nsuch as `omni.isaac.core`. In Isaac Sim 4.5, Isaac Sim extensions have been renamed\r\nto use the prefix `isaacsim`, replacing `omni.isaac`. In addition, many extensions\r\nhave been renamed and split into multiple extensions to prepare for a more modular\r\nframework that can be customized by users through the use of app templates.\r\n\r\nNotably, the following commonly used Isaac Sim extensions in Isaac Lab are renamed as follow:\r\n\r\n* `omni.isaac.cloner` --> `isaacsim.core.cloner`\r\n* `omni.isa","2025-01-30T22:56:47",{"id":216,"version":217,"summary_zh":218,"released_at":219},127113,"v1.4.1","## 👀 Overview\r\n\r\nThis release contains a set of improvements and bug fixes.\r\n\r\nMost importantly, we reverted one of the changes from the previous release (https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F966) to ensure the training throughput performance remains the same.\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fcompare\u002Fv1.4.0...v1.4.1\r\n\r\n> This is the **final release compatible with Isaac Sim 4.2**. The next release will target Isaac Sim 4.5, which introduces breaking changes that will make Isaac Lab incompatible with earlier versions of Isaac Sim.\r\n\r\n## ✨ New Features\r\n\r\n* Adds documentation and demo script for IMU sensor by @mpgussert in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1694\r\n\r\n## 🔧 Improvements\r\n\r\n* Removes deprecation for root_state_w properties and setters by @jtigue-bdai in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1695\r\n* Fixes MARL workflows for recording videos during training\u002Finferencing by @Rishi-V in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1596\r\n* Adds body tracking option to ViewerCfg by @KyleM73 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1620\r\n* Fixes the `joint_parameter_lookup` type in `RemotizedPDActuatorCfg` to support list format by @fan-ziqi in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1626\r\n* Updates pip installation documentation to clarify options by @steple in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1621\r\n* Fixes docstrings in Articulation Data that report wrong return dimension by @zoctipus in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1652\r\n* Fixes documentation error for PD Actuator by @kellyguo11 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1668\r\n* Clarifies ray documentation and fixes minor issues by @garylvov in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1717\r\n* Updates code snippets in documentation to reference scripts by @mpgussert in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1693\r\n* Adds dict conversion test for ActuatorBase configs by @mschweig in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1608\r\n\r\n## 🐛 Bug Fixes\r\n\r\n* Fixes JointAction not preserving order when using all joints by @T-K-233 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1587\r\n* Fixes event term for pushing root by setting velocity by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1584\r\n* Fixes error in Articulation where `default_joint_stiffness` and `default_joint_damping` are not correctly set for implicit actuator by @zoctipus in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1580\r\n* Fixes action reset of `pre_trained_policy_action` in navigation environment by @nicolaloi in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1623\r\n* Fixes rigid object's root com velocities timestamp check by @ori-gadot in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1674\r\n* Adds interval resampling on event manager's reset call by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1750\r\n* Corrects calculation of target height adjustment based on sensor data by @fan-ziqi in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1710\r\n* Fixes infinite loop in `repeated_objects_terrain` method  by @nicolaloi in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1612\r\n* Fixes issue where the indices were not created correctly for articulation setters by @AntoineRichard in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1660\r\n\r\n## 🤗 New Contributors\r\n\r\n* @T-K-233 made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1587\r\n* @steple made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1616\r\n* @Rishi-V made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1596\r\n* @nicolaloi made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1623\r\n* @mschweig made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1608\r\n* @AntoineRichard made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1660\r\n* @ori-gadot made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1674\r\n* @garylvov made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1717","2025-01-30T20:59:05",{"id":221,"version":222,"summary_zh":223,"released_at":224},127114,"v1.4.0","## 👀 Overview\r\n\r\nDue to a great amount of amazing updates, we are putting out one more Isaac Lab release based off of Isaac Sim 4.2. This release contains many great new additions and bug fixes, including several new environments, distributed training and hyperparameter support with Ray, new live plot feature for Manager-based environments, and more.\r\n\r\nWe will now spend more focus on the next Isaac Lab release geared towards the new Isaac Sim 4.5 release coming soon. The upcoming release will contain breaking changes in both Isaac Lab and Isaac Sim and breaks backwards compatibility, but will come with many great fixes and improvements.\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fcompare\u002Fv1.3.0...v1.4.0\r\n\r\n## ✨ New Features\r\n* Adds Factory contact-rich manipulation tasks to IsaacLab by @noseworm in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1520\r\n* Adds a Franka stacking ManagerBasedRLEnv by @peterd-NV in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1494\r\n* Adds recorder manager in manager-based environments by @nvcyc in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1336\r\n* Adds Ray Workflow: Multiple Run Support, Distributed Hyperparameter Tuning, and Consistent Setup Across Local\u002FCloud by @glvov-bdai in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1301\r\n* Adds `OperationSpaceController` to docs and tests and implement corresponding action\u002Faction_cfg classes by @ozhanozen in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F913\r\n* Adds null-space control option within `OperationSpaceController` by @ozhanozen in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1557\r\n* Adds observation term history support to Observation Manager by @jtigue-bdai in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1439\r\n* Adds live plots to managers by @pascal-roth in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F893\r\n\r\n## 🔧 Improvements\r\n* Adds documentation and example scripts for sensors by @mpgussert in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1443\r\n* Removes duplicated `TerminationsCfg` code in G1 and H1 RoughEnvCfg by @fan-ziqi in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1484\r\n* Adds option to change the clipping behavior for all Cameras and unifies the default by @pascal-roth in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F891\r\n* Adds check that no articulation root API is applied on rigid bodies by @lgulich in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1358\r\n* Adds RayCaster rough terrain base height to reward by @Andy-xiong6 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1525\r\n* Adds position threshold check for state transitions by @DorsaRoh in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1544\r\n* Adds clip range for JointAction by @fan-ziqi in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1476\r\n\r\n## 🐛 Bug Fixes\r\n* Fixes noise_model initialized in direct_marl_env by @NoneJou072 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1480\r\n* Fixes entry_point and kwargs in isaaclab_tasks README by @fan-ziqi in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1485\r\n* Fixes syntax for checking if pre-commit is installed in isaaclab.sh by @louislelay in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1422\r\n* Corrects fisheye camera projection types in spawner configuration by @command-z-z in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1361\r\n* Fixes actuator velocity limits propagation down the articulation root_physx_view by @jtigue-bdai in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1509\r\n* Computes Jacobian in the root frame inside the `DifferentialInverseKinematicsAction` class by @zoctipus in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F967\r\n* Adds transform for mesh_prim of ray caster sensor by @clearsky-mio in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1448\r\n* Fixes configclass dict conversion for torch tensors by @lgulich in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1530\r\n* Fixes error in apply_actions method in `NonHolonomicAction` action term. by @KyleM73 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1513\r\n* Fixes outdated sensor data after reset by @kellyguo11 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1276\r\n* Fixes order of logging metrics and sampling commands in command manager by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1352\r\n\r\n## 💔 Breaking Changes\r\n* Refactors pose and velocities to link frame and COM frame APIs by @jtigue-bdai in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F966\r\n\r\n## 🤗 New Contributors\r\n* @nvcyc made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1336\r\n* @peterd-NV made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1494\r\n* @NoneJou072 made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1480\r\n* @clearsky-mio made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1448\r\n* @Andy-xiong6 made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1525\r\n* @noseworm made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1520\r\n","2024-12-20T16:10:07",{"id":226,"version":227,"summary_zh":228,"released_at":229},127115,"v1.3.0","## 👀 Overview\r\n\r\nThis release will be a final release based on Isaac Sim 4.2 before the transition to Isaac Sim 4.5, which will likely contain breaking changes and no longer backwards compatible with Isaac Sim 4.2 and earlier. In this release, we introduce many features, improvements, and bug fixes, including IMU sensors, support for various types of gymnasium spaces, manager-based perception environments, and more.\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fcompare\u002Fv1.2.0...v1.3.0\r\n\r\n## ✨ New Features\r\n\r\n* Adds `IMU` sensor  by @pascal-roth in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F619\r\n* Add Camera Benchmark Tool and Allow Correct Unprojection of distance_to_camera depth image by @glvov-bdai in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F976\r\n* Creates Manager Based Cartpole Vision Example Environments by @glvov-bdai in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F995\r\n* Adds image extracted features observation term and cartpole examples for it by @glvov-bdai in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1191\r\n* Supports other gymnasium spaces in Direct workflow by @Toni-SM in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1117\r\n* Adds configuration classes for spawning different assets at prim paths by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1164\r\n* Adds a rigid body collection class by @Dhoeller19 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1288\r\n* Adds option to scale\u002Ftranslate\u002Frotate meshes in the `mesh_converter` by @pascal-roth in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1228\r\n* Adds event term to randomize gains of explicit actuators by @MoreTore in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1005\r\n* Adds Isaac Lab Reference Architecture documentation by @OOmotuyi in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1371\r\n\r\n## 🔧 Improvements\r\n\r\n* Expands functionality of FrameTransformer to allow multi-body transforms by @jsmith-bdai in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F858\r\n* Inverts SE-2 keyboard device actions (Z, X)  for yaw command by @riccardorancan in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1030\r\n* Disables backward pass compilation of warp kernels by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1222\r\n* Replaces TensorDict with native dictionary by @Toni-SM in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1348\r\n* Improves omni.isaac.lab_tasks loading time by @Toni-SM in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1353\r\n* Caches PhysX view's joint paths when processing fixed articulation tendons by @Toni-SM in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1347\r\n* Replaces hardcoded module paths with `__name__` dunder by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1357\r\n* Expands observation term scaling to support list of floats by @pascal-roth in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1269\r\n* Removes extension startup messages from the Simulation App by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1217\r\n* Adds a render config to the simulation and tiledCamera limitations to the docs by @kellyguo11 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1246\r\n* Adds Kit command line argument support by @kellyguo11 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1293\r\n* Modifies workflow scripts to generate random seed when seed=-1 by @kellyguo11 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1048\r\n* Adds benchmark script to measure robot loading by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1195\r\n* Switches from `carb` to `omni.log` for logging by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1215\r\n* Excludes cache files from vscode explorer by @Divelix in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1131\r\n* Adds versioning to the docs by @sheikh-nv in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1247\r\n* Adds better error message for invalid actuator parameters by @lgulich in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1235\r\n* Updates tested docker and apptainer versions for cluster deployment by @pascal-roth in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1230\r\n* Removes `ml_archive` as a dependency of `omni.isaac.lab` extension by @fan-ziqi in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1266\r\n* Adds a validity check for configclasses by @Dhoeller19 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1214\r\n* Ensures mesh name is compatible with USD convention in mesh converter by @fan-ziqi in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1302\r\n* Adds sanity check for the term type inside the command manager by @command-z-z in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1315\r\n* Allows configclass `to_dict` operation to handle a list of configclasses by @jtigue-bdai in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1227\r\n\r\n## 🐛 Bug Fixes\r\n\r\n* Disables replicate physics for deformable teddy lift environment by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1026\r\n* Fixes Jacobian joint indices for floating base articulations by @lorenwel in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F1033\r\n* Fixes setting the seed from CLI","2024-11-21T23:01:07",{"id":231,"version":232,"summary_zh":233,"released_at":234},127116,"v1.2.0","## 👀 Overview\r\n\r\nWe leverage the new release of Isaac Sim, 4.2.0, and bring RTX-based tiled rendering, support for multi-agent environments, and introduce many bug fixes and improvements.\r\n\r\nAdditionally, we have published an example for generating rewards using an LLM based on [Eureka](https:\u002F\u002Fgithub.com\u002Feureka-research\u002FEureka), available here https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLabEureka.\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fcompare\u002Fv1.1.0...v1.2.0\r\n\r\n##  🔆 Highlighted Features \r\n\r\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002Fb991ff62-c6b2-4042-9435-ac7ae232fb26\r\n\r\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002Fcb929b45-63b7-433c-af6c-5c19b4951b35\r\n\r\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F4ecc18e0-ec6c-455a-9a2f-88b00bae0a9f\r\n\r\n## ✨ New Features\r\n\r\n* Adds RTX-based tiled rendering. This improves the overall rendering speed and quality.\r\n* Adds the direct workflow perceptive Shadowhand Cube Repose environment `Isaac-Repose-Cube-Shadow-Vision-Direct-v0` by @kellyguo11.\r\n* Adds support for multi-agent environments with the Direct workflow, with support for MAPPO and IPPO in SKRL by @Toni-SM \r\n* Adds the direct workflow multi-agent environments `Isaac-Cart-Double-Pendulum-Direct-v0` and `Isaac-Shadow-Hand-Over-Direct-v0` by @Toni-SM \r\n* Adds throughput benchmarking scripts for the different learning workflows by @kellyguo11 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F759\r\n* Adds results for the benchmarks in the documentation [here](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fsource\u002Foverview\u002Freinforcement-learning\u002Fperformance_benchmarks.html) for different types of hardware by @kellyguo11 \r\n* Adds the direct workflow Allegro hand environment by @kellyguo11 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F709\r\n* Adds video recording to the play scripts in RL workflows by @j3soon in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F763\r\n* Adds comparison tables for the supported RL libraries [here](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fsource\u002Foverview\u002Freinforcement-learning\u002Frl_frameworks.html) by @kellyguo11 \r\n* Add APIs for deformable asset by @masoudmoghani in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F630\r\n* Adds support for MJCF converter by @qqqwan in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F957\r\n* Adds a function to define camera configs through intrinsic matrix by @pascal-roth in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F617\r\n* Adds configurable modifiers to observation manager by @jtigue-bdai in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F830\r\n* Adds the Hydra configuration system for RL training by @Dhoeller19 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F700\r\n\r\n## 🔧 Improvements \r\n\r\n* Uses PhysX accelerations for rigid body acceleration data by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F760\r\n* Adds documentation on the frames for asset data by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F742\r\n* Renames Unitree configs in locomotion tasks to match properly by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F714\r\n* Adds option to set the height of the border in the `TerrainGenerator` by @pascal-roth in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F744\r\n* Adds a cli arg to `run_all_tests.py` for testing a selected extension by @jsmith-bdai in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F753\r\n* Decouples rigid object and articulation asset classes by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F644\r\n* Adds performance optimizations for domain randomization by @kellyguo11 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F494\r\n* Allows having hybrid dimensional terms inside an observation group by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F772\r\n* Adds a flag to preserve joint order inside `JointActionCfg` action term by @xav-nal in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F787\r\n* Adds the ability to resume training from a checkpoint with rl_games by @sizsJEon in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F797\r\n* Adds windows configuration to VS code tasks by @johnBuffer in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F963\r\n* Adapts A and D button bindings in the keyboard device by @zoctipus in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F910\r\n* Uses `torch.einsum` for  quat_rotate and quat_rotate_inverse operations by @dxyy1 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F900\r\n* Expands on articulation test for multiple instances and devices by @jsmith-bdai in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F872\r\n* Adds setting of environment seed at initialization by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F940\r\n* Disables default viewport when headless but cameras are enabled by @kellyguo11 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F851\r\n* Simplifies the return type for `parse_env_cfg` method by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F965\r\n* Simplifies the if-elses inside the event manager apply method by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F948\r\n\r\n## 🐛 Bug Fixes\r\n\r\n* Fixes rendering frame delays. Rendered ","2024-09-20T17:44:31",{"id":236,"version":237,"summary_zh":238,"released_at":239},127117,"v1.1.0","## 👀 Overview\r\n\r\nWith the release of Isaac Sim 4.0 and 4.1, support for Isaac Sim 2023.1.1 has been discontinued. We strongly encourage all users to upgrade to Isaac Sim 4.1 to take advantage of the latest features and improvements. For detailed information on this upgrade, please refer to the release notes available [here](https:\u002F\u002Fdocs.omniverse.nvidia.com\u002Fisaacsim\u002Flatest\u002Frelease_notes.html).\r\n\r\nBesides the above, the Isaac Lab release brings new features and improvements, as detailed below. We thank all our contributors for their continued support.\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fcompare\u002Fv1.0.0...v1.1.0\r\n\r\n## ✨ New Features\r\n\r\n* Adds distributed multi-GPU learning support for skrl by @Toni-SM in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F574\r\n* Updates skrl integration to support training\u002Fevaluation using JAX by @Toni-SM in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F592\r\n* Adds lidar pattern for raycaster sensor by @pascal-roth in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F616\r\n* Adds support for PBS job scheduler-based clusters by @shafeef901 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F605\r\n* Adds APIs for spawning deformable meshes by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F613\r\n\r\n## 🔧 Improvements \r\n\r\n* Changes documentation color to the green theme by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F585\r\n* Fixes sphinx tabs to make them work in dark theme by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F584\r\n* Fixes VSCode settings to work with pip installation of Isaac Sim by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F628\r\n* Fixes `isaaclab` scripts to deal with Isaac Sim pip installation by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F631\r\n* Optimizes interactive scene for homogeneous cloning by @kellyguo11 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F636\r\n* Improves docker X11 forwarding documentation by @j3soon in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F685\r\n\r\n## 🐛 Bug Fixes\r\n\r\n* Reads gravity direction from simulation inside `RigidObjectData` by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F582\r\n* Fixes reference count in asset instances due to circular references by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F580\r\n* Fixes issue with asset deinitialization due to torch > 2.1 by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F640\r\n* Fixes the rendering logic regression in environments by @Dhoeller19 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F614\r\n* Fixes the check for action-space inside Stable-Baselines3 wrapper by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F610\r\n* Fixes warning message in Articulation config processing by @locoxsoco in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F699\r\n* Fixes action term in the reach environment by @masoudmoghani in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F710\r\n* Fixes training UR10 reach with RL_GAMES and SKRL by @sudhirpratapyadav in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F678\r\n* Adds event manager call to simple manage-based env by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F666\r\n\r\n## 💔 Breaking Changes\r\n\r\n* Drops official support for Isaac Sim 2023.1.1\r\n* Removes the use of body view inside the asset classes by @Mayankm96 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F643\r\n* Renames `SimulationCfg.substeps` to `SimulationCfg.render_interval` by @Dhoeller19 in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F515\r\n\r\n## ✈️ Migration Guide\r\n\r\n### Renaming of `SimulationCfg.substeps`\r\n\r\nPreviously, the users set both `omni.isaac.lab.sim.SimulationCfg.dt` and `omni.isaac.lab.sim.SimulationCfg.substeps`, which marked the physics insulation time-step and sub-steps, respectively. It was unclear whether sub-steps meant the number of integration steps inside the physics time-step `dt` or the number of physics steps inside a rendering step. \r\n\r\nSince in the code base, the attribute was used as the latter, it has been renamed to `render_interval` for clarity.\r\n\r\n### Removal of Deprecated Attributes\r\n\r\nAs notified in previous releases, we removed the classes and attributes marked as deprecated. These are as follows:\r\n\r\n* The `mdp.add_body_mass` method in the events. Please use the `mdp.randomize_rigid_body_mass` instead.\r\n* The classes `managers.RandomizationManager` and `managers.RandomizationTermCfg`. Please use the `managers.EventManager` and `managers.EventTermCfg` classes instead.\r\n* The following properties in `omni.isaac.lab.sensors.FrameTransformerData`:\r\n   * `target_rot_source` --> `target_quat_w`\r\n   * `target_rot_w` --> `target_quat_source`\r\n   * `source_rot_w` --> `source_quat_w`\r\n* The attribute `body_physx_view` from the `omni.isaac.lab.assets.Articulation` and `omni.isaac.lab.assets.RigidObject` classes. These caused confusion when used with the articulation view since the body names did not follow the same ordering.\r\n\r\n##  🤗 New Contributors\r\n\r\n* @Brayden-Zhang made their first contribution in https:\u002F\u002Fg","2024-07-26T21:06:04",{"id":241,"version":242,"summary_zh":243,"released_at":244},127118,"v1.0.0","## 👀 Overview\r\n\r\nWelcome to the first official release of Isaac Lab!\r\n\r\nBuilding upon the foundation of the [Orbit](https:\u002F\u002Fisaac-orbit.github.io\u002F) framework, we have integrated the RL environment designing workflow from [OmniIsaacGymEnvs](https:\u002F\u002Fgithub.com\u002FNVIDIA-Omniverse\u002FOmniIsaacGymEnvs). This allows users to choose a suitable [task-design approach](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fsource\u002Ffeatures\u002Ftask_workflows.html) for their applications.\r\n\r\nWhile we maintain backward compatibility with Isaac Sim 2023.1.1, we highly recommend using Isaac Lab with Isaac Sim 4.0.0 version for the latest features and improvements.\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fcompare\u002Fv0.3.1...v1.0.0\r\n\r\n## ✨ New Features\r\n\r\n* Integrated CI\u002FCD pipeline, which is triggered on pull requests and publishes the results publicly \r\n* Extended support for Windows OS platforms\r\n* Added [tiled rendered](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fsource\u002Ffeatures\u002Ftiled_rendering.html) based Camera sensor implementation. This provides optimized RGB-D rendering throughputs of up to 10k frames per second.\r\n* Added support for multi-GPU and multi-node training for the RL-Games library\r\n* Integrated APIs for environment designing (direct workflow) without relying on managers\r\n* Added implementation of delayed PD actuator model\r\n* [Added various new learning environments](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fsource\u002Ffeatures\u002Fenvironments.html) :\r\n  * Cartpole balancing using images\r\n  * Shadow hand cube reorientation\r\n  * Boston Dynamics Spot locomotion\r\n  * Unitree H1 and G1 locomotion\r\n  * ANYmal-C navigation\r\n  * Quadcopter target reaching\r\n\r\n## 🔧 Improvements \r\n\r\n* Reduced start-up time for scripts (inherited from Isaac Sim 4.0 improvements)\r\n* Added lazy buffer implementation for rigid object and articulation data. Instead of updating all the quantities at every step call, the lazy buffers are updated only when the user queries them\r\n* Added SKRL support to more environments\r\n\r\n## 💔 Breaking Change\r\n\r\nFor users coming from Orbit, this release brings certain breaking changes. Please check the migration guide for more information.\r\n\r\n## ✈️ Migration Guide\r\n\r\nPlease find detailed migration guides as follows:\r\n\r\n* [From Orbit to IsaacLab](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fsource\u002Fmigration\u002Fmigrating_from_orbit.html)\r\n* [From OmniIsaacGymEnvs to IsaacLab](https:\u002F\u002Fisaac-sim.github.io\u002FIsaacLab\u002Fsource\u002Fmigration\u002Fmigrating_from_omniisaacgymenvs.html)\r\n\r\n## 🤗 New Contributors\r\n\r\n* @abizovnuralem made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F452\r\n* @eltociear made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F460\r\n* @zoctipus made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F486\r\n* @JunghwanRo made their first contribution in https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002FIsaacLab\u002Fpull\u002F497\r\n\r\n## 🌟 Acknowledgements\r\n\r\nWe wholeheartedly thank @Mayankm96, @kellyguo11 and the Boston Dynamics AI Institute for their significant contributions to the framework.","2024-06-26T08:37:57",{"id":246,"version":247,"summary_zh":248,"released_at":249},127119,"v0.3.1","## 👀 Overview\r\n\r\nThis is an intermediate release with minor patch fixes and improvements.\r\n\r\n**Note:** This is the final release of the framework under the name \"Orbit\". The framework will soon be renamed to \"Isaac Lab\" and be moved to a [new GitHub organization](https:\u002F\u002Fgithub.com\u002Fisaac-sim\u002F). More details will be provided shortly.\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FNVIDIA-Omniverse\u002Forbit\u002Fcompare\u002Fv0.3.0...v0.3.1\r\n\r\n## ✨ New Features\r\n\r\n* Allows setting USD variants when loading prim from USD file by @lorenwel\r\n* Adds [`fix_root_link`](https:\u002F\u002Fisaac-orbit.github.io\u002Forbit\u002Fsource\u002Fhow-to\u002Fmake_fixed_prim.html) attribute to ArticulationRootPropertiesCfg by @Mayankm96\r\n\r\n## 🔧 Improvements \r\n\r\n* Removes duplicated cassie configuration in core extension by @MuhongGuo\r\n* Fixes type-hinting for articulation properties in `from_files_cfg.py` by @MuhongGuo\r\n* Adds X11 setup checks for the Docker container by @hhansen-bdai\r\n\r\n## 🐛 Bug Fixes\r\n\r\n* Fixes configclass shared references to keep compound objects independent across subclass instances by @hhansen-bdai\r\n* Fixes loading of ContactSensor when using it in an extension by @fyu-bdai\r\n* Fixes RSL-RL ONNX exporter for empirical normalization by @Nemantor \r\n\r\n##  🤗 New Contributors\r\n\r\n* @MuhongGuo made their first contribution in https:\u002F\u002Fgithub.com\u002FNVIDIA-Omniverse\u002Forbit\u002Fpull\u002F383\r\n* @lorenwel made their first contribution in https:\u002F\u002Fgithub.com\u002FNVIDIA-Omniverse\u002Forbit\u002Fpull\u002F402","2024-05-31T13:05:15",{"id":251,"version":252,"summary_zh":253,"released_at":254},127120,"v0.3.0","## 👀 Overview\r\n\r\nThis release includes various fixes and improvements to the framework. It additionally includes many new features, as listed below.\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FNVIDIA-Omniverse\u002Forbit\u002Fcompare\u002Fv0.2.0...v0.3.0\r\n\r\n## ✨ New Features\r\n\r\n* Adds terrain-aware patch sampling into the terrain generator by @nikitardn \r\n* Adds a viewer camera controller to the base environment by @farbod-farshidian\r\n* Adds animation recording for environments by @Mayankm96\r\n* Adds Franka cabinet opening environment from IsaacGymEnvs by @renezurbruegg\r\n* Adds Allegro hand cube manipulation environment from IsaacGymEnvs by @Mayankm96, @arbhardwaj98 \r\n* Adds Allegro and Shadow Hand asset configurations by @Mayankm96\r\n* Adds Kinova (Jaco2, Gen3) asset configurations by @Mayankm96 \r\n* Adds approximate torque calculation for implicit actuator by @nikitardn\r\n* Adds loading of [custom kit experience](https:\u002F\u002Fgithub.com\u002FNVIDIA-Omniverse\u002Forbit\u002Ftree\u002Fmain\u002Fsource\u002Fapps) files using AppLauncher by @Mayankm96 \r\n\r\n## 🔧 Improvements \r\n\r\n* Adds a runner script to execute all tests in the `source` directory by @jsmith-bdai\r\n* Allows loading of initialized configs in `load_cfg_from_registry` by @nikitardn\r\n* Checks default joint states are configured within limits by @Dhoeller19\r\n* Expands on unit tests for rigid object asset and terrain importer classes by @jsmith-bdai\r\n* Adds unit tests for contact sensor class by @fyu-bdai\r\n* Adds signal interrupt handle to AppLauncher by @Mayankm96\r\n* Adds ROS2 Humble to Dockerfile by @hhansen-bdai\r\n* Improved usage instructions for cluster deployment by @pascal-roth \r\n* Removes unnecessary future imports for Python 3.10 by @Mayankm96 \r\n\r\n## 🐛 Bug Fixes\r\n\r\n* Fixes running environments with a single instance by @Dhoeller19\r\n* Fixes source frame indexing in FrameTransfomer sensor by @jsmith-bdai\r\n* Fixes handling of time-out signal in RSL-RL and RL-Games wrapper by @Mayankm96\r\n* Fixes unwanted squeeze in Articulation class for 1-joint assets by @Mayankm96\r\n* Fixes shape argument ordering in `hf_terrains.random_uniform_terrain` by @nikitardn\r\n* Fixes joint and body sub-indexing for observations and rewards by @Dhoeller19\r\n* Fixes camera sensor for Isaac Sim 2023.1 update by @hhansen-bdai \r\n* Fixes imitation learning workflow for lift environment by @jsmith-bdai\r\n* Fixes apply actions method in the `NonHolonomicAction` action term class by @KyleM73\r\n* Fixes the tensor shape for the contact sensor's force matrix data by @abmoRobotics\r\n* Fixes missing max raycast distance in RayCaster sensor by @renezurbruegg\r\n* Fixes rendering of RTX sensors within the environment stepping by @Dhoeller19 \r\n\r\n## 💔 Breaking Changes\r\n\r\n* Drops support for Isaac Sim 2022.2.2 and earlier by @hhansen-bdai\r\n* Removes `compat` submodule from orbit by @Mayankm96\r\n* Removes `omni.isaac.contrib_tasks` in favor of [separate project template](https:\u002F\u002Fgithub.com\u002Fisaac-orbit\u002Forbit.ext_template) by @nburger-bdai \r\n* Changes link names ordering in articulation to follow PhysX by @Mayankm96\r\n* Renames `RandomizationManager` to `EventManager` by @pascal-roth\r\n* Renames some terms in `omni.isaac.orbit.envs.mdp` to avoid confusion by @Mayankm96 \r\n\r\n## ✈️ Migration Guide\r\n\r\n### Renaming of Randomization Manager to Event Manager\r\n\r\nWhile the randomization manager referred to all possible \" randomizations \" in the environment, many users felt that the triggering\r\nof the terms was not apparent. Additionally, for non-RL use-cases, randomization seemed like a misnomer. Hence, we renamed the class to be called event manager.\r\n\r\nReplace the following import and usage:\r\n\r\n```python\r\nfrom omni.isaac.orbit.managers import RandomizationTermCfg as RandTerm\r\n\r\n@configclass\r\nclass RandomizationCfg:\r\n    \"\"\"Configuration for randomization.\"\"\"\r\n\r\n    reset_base = RandTerm(\r\n        func=mdp.reset_root_state_uniform,\r\n        mode=\"reset\",\r\n        params={\"pose_range\": {}, \"velocity_range\": {}},\r\n    )\r\n\r\n@configclass\r\nclass MyEnvCfg:\r\n\r\n     randomization: RandomizationCfg = RandomizationCfg()\r\n```\r\n\r\nwith the following:\r\n\r\n```python\r\nfrom omni.isaac.orbit.managers import EventTermCfg as EventTerm\r\n\r\n@configclass\r\nclass EventCfg:\r\n    \"\"\"Configuration for events.\"\"\"\r\n\r\n    reset_base = EventTerm(\r\n        func=mdp.reset_root_state_uniform,\r\n        mode=\"reset\",\r\n        params={\"pose_range\": {}, \"velocity_range\": {}},\r\n    )\r\n\r\n@configclass\r\nclass MyEnvCfg:\r\n\r\n     events: EventCfg = EventCfg()\r\n```\r\n\r\n### Renaming of MDP terms\r\n\r\nWe observed that several of the MDP terms had close to overlapping names, such as the reward term `joint_pos_limits` and termination term `joint_pos_limit`. To avoid errors, we decided to make the term names clearer.\r\n\r\nPlease note the following changes in MDP terms:\r\n\r\n  * Observation: `joint_pos_norm` -> `joint_pos_limit_normalized`\r\n  * Event: `add_body_mass` -> `randomize_rigid_body_mass`\r\n  * Termination: `base_height` -> `root_height_below_minimum`\r\n  * Termination: `joint_pos_limit` -> `joint_pos_out_of_limit`\r\n","2024-04-17T20:01:45",{"id":256,"version":257,"summary_zh":258,"released_at":259},127121,"v0.2.0","## 👀 Overview\r\n\r\nThis release refactors Orbit APIs to make them more modular and performance-efficient. It also introduces various new functionalities, such as batched sensors and managers handling randomization, termination, action spaces, and curriculum.\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FNVIDIA-Omniverse\u002Forbit\u002Fcompare\u002Fv0.1.0...v0.2.0\r\n\r\n## ✨ New Features\r\n\r\n* Procedural terrain generation using height-fields and `trimesh` library\r\n* Sensors for contact sensing, frame transformations, and GPU-based ray-casting\r\n* [Scene manager](https:\u002F\u002Fisaac-orbit.github.io\u002Forbit\u002Fsource\u002Fapi\u002Forbit\u002Fomni.isaac.orbit.scene.html#module-omni.isaac.orbit.scene) to handle a collection of assets and sensors\r\n* Managers to handle various aspects of environment designing - action space, observation space, randomization, termination, rewards, and curriculum\r\n* Spawning support to load meshes in different formats (OBJ, FBX, STL) and URDF files into simulation directly\r\n* Support for docker and cluster-based deployments\r\n* [New environments](https:\u002F\u002Fisaac-orbit.github.io\u002Forbit\u002Fsource\u002Ffeatures\u002Fenvironments.html) for flat and rough terrain locomotion from [`legged_gym`](https:\u002F\u002Fgithub.com\u002Fleggedrobotics\u002Flegged_gym)\r\n\r\n## 🔧 Improvements \r\n\r\n* Merged the previous type-specific static markers into a unified [`VisualizationMarkers`](https:\u002F\u002Fisaac-orbit.github.io\u002Forbit\u002Fsource\u002Fapi\u002Forbit\u002Fomni.isaac.orbit.markers.html) class\r\n* Switched to using PhysX interfaces directly instead of Isaac Sim interface classes for performance\r\n* Migrated the `gym.Env` class from the OpenAI Gym 0.21 version to the Gymnasium v0.29 version\r\n\r\n## 💔 Breaking Changes\r\n\r\nThis release is incompatible with v0.1.0 due to heavy refactoring and changes. We recommend users revisit the tutorials to familiarize themselves with the new APIs. We apologize for any inconvenience caused but believe the enhancements in this release justify the transition.\r\n\r\n##  🤗 New Contributors\r\n\r\n* @kouroshD\r\n* @Toni-SM\r\n* @pascal-roth\r\n* @nikitardn \r\n* @Dhoeller19 \r\n* @jsmith-bdai \r\n* @AutonomousHansen \r\n* @farbod-farshidian ","2024-02-10T12:22:32"]