{"success":true,"data":{"author":"University of California (T-Rex authors)","bucketAccessType":"public","bucketName":"zekaiwang/trex_dataset","bucketPrefix":"","category":"Robotics","cloudProvider":"huggingface","createdAt":"2026-11-14T14:15:10.822Z","description":"T-Rex is a large tactile-reactive bimanual manipulation dataset collected by teleoperation on a Dexmate Vega-1 dual-arm robot fitted with two Sharpa Wave dexterous hands, each with five fingertip tactile sensors (arXiv:2606.17055). Manus gloves and VIVE trackers drive the robot; observations come from a head-mounted ZED X Mini camera and two wide-view ZED X One S wrist cameras, recorded at 640x360. Per fingertip the dataset stores the raw tactile sensor image and an estimated deformation map, both as video, and the estimated 6-axis net wrench. The card reports 5,464 episodes, 5,473,459 frames (about 50 hours at 30 fps), 5,370 language-annotated trajectories, 22 motor primitives and 207 objects. It is a LeRobot v3.0 dataset with Parquet under data/, MP4 under videos/ and an episodes_preview.parquet index. Measured 2026-11-14 at about 1.55 TB. Load with lerobot's LeRobotDataset, pandas or pyarrow; quickstart code is in the T-Rex GitHub repo. Useful for tactile-conditioned policies and contact-rich dexterous manipulation. Released under the MIT licence, copyright the Regents of the University of California; cite the T-Rex paper.","displayTitle":"T-Rex: Tactile-Reactive Bimanual Dexterous Manipulation Dataset","id":"dcc77616-cadc-5d31-b506-4487e9ef8171","license":"MIT","licenseEvidenceUrl":"https://huggingface.co/datasets/zekaiwang/trex_dataset","objectCountEstimate":7903,"probeStatus":"ok","referenceUrl":"https://arxiv.org/abs/2606.17055","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"zekaiwang/trex_dataset\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"zekaiwang/trex_dataset\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=zekaiwang, license:mit, size_categories:1M<n<10M, format:parquet, downloads=70770"],"format":"mixed","generatedAt":"2026-11-14T14:15:15.289Z","layout":{"example":"data/chunk-000/file-000.parquet","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"data/chunk-000/file-000.parquet","truncatedAt":512}],"sampleObjects":["assets/dataset_analysis.png","assets/dataset_overview.png","assets/hardware_setup.png","data/chunk-000/file-000.parquet","data/chunk-000/file-001.parquet"],"sidecars":[{"bytes":"MIT License\n\nCopyright (c) 2026 The Regents of the University of California\n\nPermission is hereby granted, free of charge, to any person obtaining a copy\nof this software and associated documentation files (the \"Software\"), to deal\nin the Software without restriction, including without limitation the rights\nto use, copy, modify, merge, publish, distribute, sublicense, and/or sell\ncopies of the Software, and to permit persons to whom the Software is\nfurnished to do so, subject to the following conditions:\n\nThe above copyright notice and this permission notice shall be included in all\ncopies or substantial portions of the Software.\n\nTHE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\nIMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\nFITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE\nAUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\nLIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,\nOUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE\nSOFTWARE.\n","encoding":"utf8","key":"LICENSE","name":"LICENSE","size":1100,"truncated":false},{"bytes":"---\nlicense: mit\nlanguage:\n- en\ntask_categories:\n- robotics\ntags:\n- LeRobot\n- robotics\n- manipulation\n- tactile\n- bimanual\n- dexterous-manipulation\npretty_name: T-Rex Dataset\nsize_categories:\n- 1M<n<10M\nconfigs:\n- config_name: frames\n  data_files:\n  - split: train\n    path: data/chunk-*/file-*.parquet\n  default: true\n- config_name: episodes\n  data_files:\n  - split: train\n    path: episodes_preview.parquet\n---\n\n# T-Rex Dataset\n\nA large-scale, tactile-reactive bimanual manipulation dataset, collected via teleoperation on a\nDexmate Vega-1 robot with two Sharpa Wave dexterous hands. Stored as a\n[LeRobotDataset v3.0](https://github.com/huggingface/lerobot).\n\n[🌐 Project Page](https://tactile-rex.github.io/) · [✍️ Paper (arXiv)](https://arxiv.org/abs/2606.17055) · [💻 Code (T-Rex)](https://github.com/ZhuoyangLiu2005/T-Rex) · [🚀 Dataset Quickstart](https://github.com/ZhuoyangLiu2005/T-Rex/tree/main/dataset_quickstart) · [📓 Colab notebook](https://colab.research.google.com/github/ZhuoyangLiu2005/T-Rex/blob/main/dataset_quickstart/quickstart.ipynb)\n\n<p align=\"center\">\n  <img src=\"assets/dataset_overview.png\" width=\"100%\">\n  <br>\n  <em>One episode from each of 20 motor primitives (head-camera view, cropped to the workspace), each with a different object.</em>\n</p>\n\n<p align=\"center\">\n  <img src=\"assets/hardware_setup.png\" width=\"100%\">\n  <br>\n  <em>Teleoperation setup: Manus gloves + VIVE trackers drive the bimanual Dexmate Vega-1 with two Sharpa Wave hands; observations come from a head-mounted ZED X Mini camera and two wide-view ZED X One S wrist cameras.</em>\n</p>\n\n## At a glance\n\n| | |\n|---|---|\n| Robot | Dexmate Vega-1 dual-arm (7 actuated joints per arm) + 2× Sharpa Wave dexterous hands (5 fingertip tactile sensors each) |\n| Modalities | 3 RGB cameras · 10 raw tactile + 10 deformation tactile videos · 6-axis fingertip wrenches · joint states/targets |\n| Episodes | 5,464 |\n| Frames | 5,473,459 (~50 hours @ 30 fps) |\n| Tasks | 5,370 language-annotated trajectories · 22 motor primitives · 207 objects |\n| Format | LeRobotDataset v3.0 (`codebase_version: v3.0`) |\n\n## Composition\n\n<p align=\"center\">\n  <img src=\"assets/dataset_analysis.png\" width=\"100%\">\n  <br>\n  <em>Object categories, episodes per motor primitive, and per-object episode counts.</em>\n</p>\n\n## Collection setup\n\nThe full teleoperation/hardware stack used to collect the dataset is open-sourced in\n[`hardware_code/`](https://github.com/ZhuoyangLiu2005/T-Rex/tree/main/hardware_code) in the T-Rex repo.\n\n- **Robot.** The Dexmate Vega-1 is a dual-arm mobile robot with 7 actuated joints per arm, here\n  equipped with two Sharpa Wave dexterous hands. During collection the wheels, torso, and head\n  joints are fixed; only the 14 arm joints and the two hands are actuated.\n- **Cameras.** A head-mounted ZED X Mini stereo camera (its left monocular RGB stream is recorded)\n  plus two wide-view ZED X One S monocular RGB cameras mounted on the wrists, posed so the head\n  camera covers the full reachable workspace while the wrist cameras keep the fingers visible\n  without significant palm occlusion. All three streams are recorded at 640×360.\n- **Tactile.** Each hand carries five fingertip tactile sensors. Per sensor, the dataset stores the\n  raw sensor image and the estimated deformation map (both as video), and the estimated 6-axis net\n  wrench (`observation.tactile_force`).\n- **Teleoperation.** Manus gloves capture fingertip positions relative to the hand base, retargeted\n  to the Sharpa Wave hands with the manufacturer's differential-inverse-kinematics package\n  (Pinocchio + CasADi). Two VIVE trackers provide SE(3) wrist poses, converted to arm joint\n  commands via differential inverse kinematics ([Pink](https://github.com/stephane-caron/pink)),\n  low-pass filtered, and tracked by the manufacturer's low-level cascade PID controller. A 30 Hz\n  high-level thread records observations and joint-space targets while asynchronously updating a\n  300 Hz low-level control thread — the dataset's 30 fps matches the high-level loop, and `action`\n  holds its 30 Hz joint-","encoding":"utf8","key":"README.md","name":"README.md","size":10043,"truncated":true}],"sourceObjectKey":"data/chunk-000/file-000.parquet","starterCell":"from datasets import load_dataset\nds = load_dataset(\"zekaiwang/trex_dataset\")\nprint(ds)","starterSamplePath":"assets/dataset_analysis.png","totalObjects":7903},"summary":"T-Rex is a large tactile-reactive bimanual manipulation dataset collected by teleoperation on a Dexmate Vega-1 dual-arm robot fitted with two Sharpa Wave dexterous hands, each…","title":"T-Rex: Tactile-Reactive Bimanual Dexterous Manipulation Dataset","updatedAt":"2026-11-14T17:28:01.289Z","visibility":"public","websiteUrl":"https://tactile-rex.github.io/"}}