[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-80587":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":8,"htmlUrl":8,"language":9,"languages":8,"totalLinesOfCode":8,"stars":10,"forks":11,"watchers":12,"openIssues":13,"contributorsCount":14,"subscribersCount":14,"size":14,"stars1d":12,"stars7d":15,"stars30d":16,"stars90d":14,"forks30d":14,"starsTrendScore":17,"compositeScore":18,"rankGlobal":8,"rankLanguage":8,"license":19,"archived":20,"fork":20,"defaultBranch":21,"hasWiki":20,"hasPages":20,"topics":22,"createdAt":8,"pushedAt":8,"updatedAt":23,"readmeContent":24,"aiSummary":25,"trendingCount":14,"starSnapshotCount":14,"syncStatus":13,"lastSyncTime":26,"discoverSource":27},80587,"SUGAR","tianshuwu\u002FSUGAR","tianshuwu",null,"Python",86,5,1,2,0,14,29,11,2.33,"MIT License",false,"main",[],"2026-06-12 02:04:04","\u003Ch1 align=\"center\">\n  🍬 SUGAR: A Scalable Human-Video-Driven Generalizable Humanoid Loco-Manipulation Learning Framework\n\u003C\u002Fh1>\n\n\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Ftianshuwu.github.io\u002F\" target=\"_blank\">\u003Cstrong>Tianshu Wu\u003C\u002Fstrong>\u003C\u002Fa>\u003Csup>1*\u003C\u002Fsup> ·\n  Xiangqi Kong\u003Csup>2*\u003C\u002Fsup> ·\n  \u003Ca href=\"https:\u002F\u002Fyuechen0614.github.io\u002F\" target=\"_blank\">\u003Cstrong>Yue Chen\u003C\u002Fstrong>\u003C\u002Fa>\u003Csup>1*\u003C\u002Fsup> ·\n  Qize Yu\u003Csup>1\u003C\u002Fsup> ·\n  \u003Ca href=\"https:\u002F\u002Falvinyh.github.io\u002F\" target=\"_blank\">\u003Cstrong>Hang Ye\u003C\u002Fstrong>\u003C\u002Fa>\u003Csup>1\u003C\u002Fsup> ·\n  Jia Li\u003Csup>1\u003C\u002Fsup> ·\n  \u003Ca href=\"https:\u002F\u002Fcfcs.pku.edu.cn\u002Fwangyizhou\u002F\" target=\"_blank\">\u003Cstrong>Yizhou Wang\u003C\u002Fstrong>\u003C\u002Fa>\u003Csup>1\u003C\u002Fsup> ·\n  \u003Ca href=\"https:\u002F\u002Fzsdonghao.github.io\u002F\" target=\"_blank\">\u003Cstrong>Hao Dong\u003C\u002Fstrong>\u003C\u002Fa>\u003Csup>1,✉\u003C\u002Fsup>\n\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n  \u003Csup>1\u003C\u002Fsup> Center on Frontiers of Computing Studies, School of Computer Science, Peking University\n  \u003Cbr>\n  \u003Csup>2\u003C\u002Fsup> School of Computer Science and Engineering, Beihang University\n\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n  \u003Csup>*\u003C\u002Fsup> Equal Contribution    \n  \u003Csup>✉\u003C\u002Fsup> Corresponding Author\n\u003C\u002Fp>\n\n\n\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Farxiv.org\u002Fabs\u002F2605.20373\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPaper-arXiv-red\">\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Ftianshuwu.github.io\u002Fsugar-humanoid\u002F\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDemo-Website-blue\">\n  \u003C\u002Fa>\n  \u003Ca href=\".\u002FLICENSE\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-MIT-green\">\n  \u003C\u002Fa>\n\u003C\u002Fp>\n\n\n\n## 📌 Overview\nSUGAR is a scalable humanoid loco-manipulation project built upon the IsaacLab manager-based framework. Given third-person videos of human-object interactions, it learns generalizable and deployable humanoid autonomous policies, enabling humanoid robots to solve challenging loco-manipulation tasks in the real-world.\n\u003Cp align=\"center\">\n  \u003Cimg src=\"assets\u002Fmethod_overview.png\" width=\"100%\">\n\u003C\u002Fp>\n\n\n\n\n## 📌 Installation\n1. clone repository and create a conda environment\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Ftianshuwu\u002FSUGAR.git\ncd SUGAR\nconda create -n sugar python=3.11\nconda activate sugar\n```\n\n2. install isaacsim\n```bash\npip install isaacsim[all,extscache]==5.1.0 --extra-index-url https:\u002F\u002Fpypi.nvidia.com\n```\n3. install isaaclab\n```bash\ncd ..\ngit clone git@github.com:isaac-sim\u002FIsaacLab.git\ncd IsaacLab\ngit checkout v2.3.0\npip install flatdict==4.0.1 --no-build-isolation\n.\u002Fisaaclab.sh --install rsl_rl\n```\n4. install sugar\n```bash\ncd ..\u002FSUGAR\npip install -e source\u002Fsugar_rl\npip install -e source\u002Fsugar_il\n# for 5090:  pip install torch==2.8.0 torchvision==0.23.0 torchaudio==2.8.0 --index-url https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcu128\n```\n\n5. download data\n```bash\npip install -U gdown\npython -m gdown 1AIJWqS5rFGl5u2Qq6jCCTHKdh51SX2Sc     # 400MB\nunzip data.zip\nrm data.zip\npython -m gdown 1wXNAjNMrfV0e-d2pQ6m9dm4xrG5lSoyD     # 50MB\nunzip descriptions.zip\nrm descriptions.zip\npython -m gdown 1Uc2SPPVvTboEgw4Scyuz3TmzNKDg-dx-     # 250MB\nunzip demo_ckpts.zip\nrm demo_ckpts.zip\n```\n\n\n## 📌 Run SUGAR\n### Inference\n```bash\n# Optional task: CarryBox, KickBox, PushBox, SitChair, StandBottle, PickBottle\n# bash inference.sh TASK_NAME (optional TRACKER_CKPT) (optional GENERATOR_CKPT)\nbash inference.sh CarryBox \n```\n### Train\n```bash\n# Optional task: CarryBox, KickBox, PushBox, SitChair, StandBottle, PickBottle\n# bash train.sh TASK_NAME (optional EXP_NAME)\nbash train.sh CarryBox  \n```\n\n## 📌 TODO List\n- [x] Release inference demo and checkpoints\n- [x] Release the complete training pipeline, including refiner, tracker, and generator\n- [x] Release processed data for all six tasks\n- [ ] Release the data processing pipeline from RGB-D human videos to training data\n- [ ] Release the sim-to-sim pipeline\n\n## 📌 Acknowledgements\nThis code implementation is based on these excellent open-source projects, thanks to:\n* **[unitree_rl_lab](https:\u002F\u002Fgithub.com\u002Funitreerobotics\u002Funitree_rl_lab) & [beyondmimic](https:\u002F\u002Fgithub.com\u002FHybridRobotics\u002Fwhole_body_tracking)**: Serves as the codebase for `sugar_rl`.\n* **[dexgraspvla](https:\u002F\u002Fgithub.com\u002FPsi-Robot\u002FDexGraspVLA)**: Serves as the codebase for `sugar il`.\n\n\n## 📌 Citation\n```bash\n@article{wu2026sugar,\n  title={SUGAR: A Scalable Human-Video-Driven Generalizable Humanoid Loco-Manipulation Learning Framework},\n  author={Wu, Tianshu and Kong, Xiangqi and Chen, Yue and Yu, Qize and Ye, Hang and Li, Jia and Wang, Yizhou and Dong, Hao},\n  journal={arXiv preprint arXiv:2605.20373},\n  year={2026}\n}\n```","SUGAR是一个基于IsaacLab框架构建的可扩展人形机器人操控学习项目，通过分析第三人称视角下的人与物体互动视频，学习并生成适用于真实世界中复杂操控任务的通用型自主策略。其核心功能包括从人类行为视频中提取有用信息，并利用这些数据训练人形机器人执行复杂的移动和操作任务。技术上，SUGAR采用了先进的强化学习算法以及模拟环境下的训练机制来确保模型的有效性和泛化能力。该项目特别适合于需要开发高级别自动化解决方案的研究机构或企业使用，在服务机器人、工业自动化等领域具有广泛的应用前景。","2026-06-11 04:01:19","CREATED_QUERY"]