{"success":true,"data":{"author":"NVIDIA Corporation","bucketAccessType":"public","bucketName":"nvidia/Arena-GR1-Manipulation-Task","bucketPrefix":"","category":"Robotics","cloudProvider":"huggingface","createdAt":"2026-11-21T22:15:32.757Z","description":"Simulated humanoid manipulation trajectories from NVIDIA for the Fourier GR1 humanoid robot opening a microwave door in the IsaacLab-Arena environment, used to post-train the GR00T N1.5 vision-language-action model. Ten human demonstrations were collected in Isaac Lab with a depth camera and keyboard, then expanded to 50 demonstrations at 50 Hz with the MimicGen trajectory generation framework. The repository has the annotated 10-demo HDF5 file, the generated 50-demo HDF5 file, and a GR00T-LeRobot conversion with Parquet episodes (36-DoF joint-position actions, 54-DoF joint-state observations, timestamps and language task indices) and 512x512 first-person RGB MP4 videos, plus episodes.jsonl, tasks.jsonl, modality.json and info.json metadata. About 5 GB (measured 2026-11-21). This is the original microwave task release; the later v3 task set is listed separately. Load the HDF5 with h5py or the LeRobot folder with the lerobot library. NVIDIA releases it under CC-BY-4.0 for commercial use; cite MimicGen (Mandlekar et al., CoRL 2023) and NVIDIA IsaacLab-Arena.","displayTitle":"IsaacLab-Arena GR1 Humanoid Microwave-Opening Task: 10 Teleoperated and 50 MimicGen Demonstrations (NVIDIA, v1)","huggingFaceUrl":"https://huggingface.co/datasets/nvidia/Arena-GR1-Manipulation-Task","id":"0cbdb0f3-932a-530c-851d-eccfb2a29895","license":"CC-BY-4.0","licenseEvidenceUrl":"https://huggingface.co/datasets/nvidia/Arena-GR1-Manipulation-Task","objectCountEstimate":108,"probeStatus":"ok","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"nvidia/Arena-GR1-Manipulation-Task\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"nvidia/Arena-GR1-Manipulation-Task\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=nvidia, license:cc-by-4.0, size_categories:1K<n<10K, format:parquet, downloads=495"],"format":"mixed","generatedAt":"2026-11-21T22:31:22.494Z","layout":{"example":"arena_gr1_manipulation_dataset_annotated.hdf5","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"arena_gr1_manipulation_dataset_annotated.hdf5","truncatedAt":512}],"sampleObjects":["arena_gr1_manipulation_dataset_annotated.hdf5","arena_gr1_manipulation_dataset_generated.hdf5","lerobot/data/chunk-000/episode_000000.parquet","lerobot/data/chunk-000/episode_000001.parquet","lerobot/data/chunk-000/episode_000002.parquet"],"sidecars":[{"bytes":"---\nlicense: cc-by-4.0\ntask_categories:\n- robotics\ntags:\n- robotics\n---\n## Dataset Description:\n\nThe Arena-GR1-Manipulation-Task dataset is multimodal collections of trajectories generated in Isaac Lab. It supports humanoid (GR1) manipulation task in IsaacLab-Arena environment. Each entry provides the full context (state, vision, language, action) needed to train and evaluate generalist robot policies for opening microwave task.\n\n| Dataset Name          | # Trajectories |\n|-----------------------|----------------|\n| GR1 Manipulation Task | 50             |\n\nThis dataset is ideal for behavior cloning, policy learning, and generalist robotic manipulation research. It has been for post-training GR00T N1.5 model.\n\nThis dataset is ready for commercial use.\n\n## Dataset Owner\nNVIDIA Corporation\n\n## Dataset Creation Date:\n10/10/2025\n\n## License/Terms of Use:\nThis dataset is governed by the Creative Commons Attribution 4.0 International License (CC-BY-4.0).\n\n## Intended Usage:\nThis dataset is intended for:\n\n- Training robot manipulation policies using behavior cloning.\n- Research in generalist robotics and task-conditioned agents.\n- Sim-to-real / Sim-to-Sim transfer studies.\n\n## Dataset Characterization:\n### Data Collection Method\n\n- Automated\n- Automatic/Sensors\n- Synthetic\n  \n10 human teleoperated demonstrations are collected through a depth camera and keyboard in Isaac Lab. All 50 demos are generated automatically using a synthetic motion trajectory generation framework, Mimicgen [1]. Each demo is generated at 50 Hz.\n\n### Labeling Method\n\nNot Applicable\n\n## Dataset Format:\nWe provide a few dataset files, including\n\n- a human-annoated 10 demonstrations in HDF5 dataset file (`arena_gr1_manipulation_dataset_annotated.hdf5`)\n- a Mimic-generated 50 demonstrations in HDF5 dataset file (`arena_gr1_manipulation_dataset_generated.hdf5`)\n- a GR00T-Lerobot formatted dataset converted from the Mimic-generated HDF5 dataset file (`lerobot`)\n\nEach demo in GR00T-Lerobot datasets consists of a time-indexed sequence of the following modalities:\n\n### Actions\n- action (FP64): joint desired positions for all body joints (36 DoF)\n\n### Observations\n- observation.state (FP64): joint positions for all body joints (54 DoF)\n\n### Task-specific\n- timestamp (FP64): simulation time in seconds of each recorded data entry.\n- annotation.human.action.task_description (INT64): index referring to the language instruction recorded in the metadata\n- annotation.human.action.valid (INT64): index indicating validity of annotaion recorded in the metadata\n- episode_index (INT64): index indicating the order of each demo\n- task_index (INT64): index used in multi-task data loader. Not applicable to Gr00t-N1 post training, always set to 0.\n\n\n### Videos\n- 512 x 512 RGB videos in mp4 format from first-person-view camera\n\nIn additional, a set of metadata describing the followings is provided,\n- `episodes.jsonl` contains a list of all the episodes in the entire dataset. Each episode contains a list of tasks and the length of the episode.\n- `tasks.jsonl` contains a list of all the tasks in the entire dataset.\n- `modality.json` contains the modality configuration.\n- `info.json` contains the dataset information.\n\n\n## Dataset Quantification:\n\n### Record Count\n\n#### GR1 Manipulation Task\n- Number of demonstrations/trajectories: 50\n- Number of RGB videos: 50\n\n### Total Storage\n\n5.16 GB\n\n## Ethical Considerations:\nNVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. When downloaded or used in accordance with our terms of service, developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse. \n\nPlease report security vulnerabilities or NVIDIA AI Concerns [here](https://www.nvidia.com/en-us/support/submit-security-vulnerability/).\n\n## Reference(s):\n[1] @inproceedings{mandlekar2023mimicgen,\n    title={MimicGen: A Data Generation System for Scalable Robot Learning using ","encoding":"utf8","key":"README.md","name":"README.md","size":4352,"truncated":true}],"sourceObjectKey":"arena_gr1_manipulation_dataset_annotated.hdf5","starterCell":"from huggingface_hub import hf_hub_download\npath = hf_hub_download(repo_id=\"nvidia/Arena-GR1-Manipulation-Task\", filename=\"arena_gr1_manipulation_dataset_annotated.hdf5\", repo_type=\"dataset\")\nprint(path)","starterSamplePath":"arena_gr1_manipulation_dataset_annotated.hdf5","totalObjects":108},"summary":"Simulated humanoid manipulation trajectories from NVIDIA for the Fourier GR1 humanoid robot opening a microwave door in the IsaacLab-Arena environment, used to post-train the…","title":"IsaacLab-Arena GR1 Humanoid Microwave-Opening Task: 10 Teleoperated and 50 MimicGen Demonstrations (NVIDIA, v1)","updatedAt":"2026-11-22T01:11:49.043Z","visibility":"public","websiteUrl":"https://huggingface.co/datasets/nvidia/Arena-GR1-Manipulation-Task"}}