[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-physical-ai-datasets":3,"mdc-a6iwr7-key":37,"related-repo-nvidia-physical-ai-datasets":8341,"related-org-nvidia-physical-ai-datasets":8409},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":26,"repoUrl":27,"updatedAt":28,"license":29,"forks":30,"topics":31,"repo":32,"sourceUrl":35,"mdContent":36},"physical-ai-datasets","find and download NVIDIA Physical AI datasets","Use when the user wants to find, download, or pick a NVIDIA Physical AI dataset on Hugging Face for autonomous-vehicle, robotics, spatial intelligence, manipulation, or neural-reconstruction workflows. Catalog of every dataset under huggingface.co\u002Fnvidia with the `PhysicalAI-` prefix, organised by domain (AV, Robotics-Manipulation, Robotics-GR00T, Robotics-mindmap, Robotics-NuRec, Spatial Intelligence, Grasping, Healthcare, Sim-Ready, Material properties), with per-dataset size, format, gating, license, and the downstream sibling skill (`ncore`, `nre`, `asset-harvester`, `nurec-fixer`) or upstream tool (Isaac Sim, CARLA, Isaac-GR00T, Cosmos-*) that consumes it. Do NOT use as a runtime — it routes you elsewhere. Trigger keywords: nvidia physical ai dataset, PhysicalAI- dataset, hf nvidia dataset, NCore dataset, NuRec dataset, GR00T dataset, GraspGen, SimReady, Cosmos-Drive-Dreams, Lyra SDG, Open-H-Embodiment, huggingface-cli download, physical_ai_av, dataset gated, RDS-HQ.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},"nvidia","NVIDIA","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fnvidia.png",[12,16,19,22,25],{"name":13,"slug":14,"type":15},"Robotics","robotics","tag",{"name":17,"slug":18,"type":15},"Datasets","datasets",{"name":20,"slug":21,"type":15},"Hugging Face","hugging-face",{"name":23,"slug":24,"type":15},"Machine Learning","machine-learning",{"name":9,"slug":8,"type":15},16,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fnurec-skills","2026-07-14T05:32:33.182445","CC-BY-4.0 AND Apache-2.0",4,[],{"repoUrl":27,"stars":26,"forks":30,"topics":33,"description":34},[],"Agent skills for Neural Reconstruction Engine","https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fnurec-skills\u002Ftree\u002FHEAD\u002Fskills\u002Fphysical-ai-datasets","---\nname: physical-ai-datasets\ndescription: >-\n  Use when the user wants to find, download, or pick a NVIDIA Physical\n  AI dataset on Hugging Face for autonomous-vehicle, robotics, spatial\n  intelligence, manipulation, or neural-reconstruction workflows.\n  Catalog of every dataset under huggingface.co\u002Fnvidia with the\n  `PhysicalAI-` prefix, organised by domain (AV, Robotics-Manipulation,\n  Robotics-GR00T, Robotics-mindmap, Robotics-NuRec, Spatial\n  Intelligence, Grasping, Healthcare, Sim-Ready, Material properties),\n  with per-dataset size, format, gating, license, and the downstream\n  sibling skill (`ncore`, `nre`, `asset-harvester`, `nurec-fixer`) or\n  upstream tool (Isaac Sim, CARLA, Isaac-GR00T, Cosmos-*) that\n  consumes it. Do NOT use as a runtime — it routes you elsewhere.\n  Trigger keywords: nvidia physical ai dataset, PhysicalAI- dataset,\n  hf nvidia dataset, NCore dataset, NuRec dataset, GR00T dataset,\n  GraspGen, SimReady, Cosmos-Drive-Dreams, Lyra SDG,\n  Open-H-Embodiment, huggingface-cli download, physical_ai_av,\n  dataset gated, RDS-HQ.\nversion: \"0.1.0\"\ntools:\n  - Shell\n  - Read\n  - Write\nlicense: CC-BY-4.0 AND Apache-2.0\ndependencies:\n  - bash\n  - python\n  - huggingface_hub\ncompatibility: >-\n  All listed datasets are hosted on Hugging Face and require\n  `huggingface_hub` (CLI: `huggingface-cli` \u002F new `hf` shim) plus a HF\n  user access token. Several datasets are GATED — they need explicit\n  license acceptance on the HF dataset page before any token can pull\n  them. Storage ranges from \u003C1 GB to >100 TB (PhysicalAI-Autonomous-Vehicles, 133 TB).\nmetadata:\n  author: NVIDIA Physical AI\n  tags:\n    - dataset-catalog\n    - physical-ai\n    - huggingface\n    - autonomous-vehicles\n    - robotics\n  hf_org: https:\u002F\u002Fhuggingface.co\u002Fnvidia\n  hf_collection: https:\u002F\u002Fhuggingface.co\u002Fcollections\u002Fnvidia\u002Fphysical-ai\n  cosmos_dataset_search: https:\u002F\u002Fbuild.nvidia.com\u002Fnvidia\u002Fcosmos-dataset-search\n  pai_av_toolkit: https:\u002F\u002Fgithub.com\u002FNVlabs\u002Fphysical_ai_av\n---\n\n# NVIDIA Physical AI Datasets (Hugging Face)\n\n## Purpose\n\nHelp the agent find, evaluate, and download an NVIDIA `PhysicalAI-*`\ndataset on Hugging Face that fits the user's downstream task —\nautonomous-vehicle reconstruction, robotics manipulation, GR00T\npost-training, spatial-intelligence research, grasping, or sim-ready\ncontent — and then hand off to the sibling skill (`ncore`, `nre`,\n`asset-harvester`, `nurec-fixer`) or upstream NVIDIA tool that\nactually consumes it.\n\n**Use this skill when:** the user asks \"is there an NVIDIA dataset\nfor X?\", \"where do I get NCore \u002F NuRec \u002F GR00T sample data?\", or is\nshopping the Hugging Face NVIDIA org for `PhysicalAI-*` collections.\n\n**Do NOT use this skill when:**\n\n- The user already knows the dataset and just wants to run a\n  pipeline — jump straight to the consuming skill.\n- The user needs a non-NVIDIA dataset (Waymo, nuScenes, KITTI, …) —\n  this catalog is NVIDIA-only.\n- The user wants to train Cosmos \u002F GR00T \u002F Isaac Sim itself — that's\n  the upstream tool's job, not this catalog's.\n\n## Overview\n\nCatalog of NVIDIA's open Physical AI dataset family on Hugging Face.\nPick by **task** (Section 2 § lookup table) or **family** (Sections 3–10).\nEvery entry lists: dataset path, size, format, license, gating, and the\ndownstream skill in this repo that consumes it.\n\n> Source of truth: \u003Chttps:\u002F\u002Fhuggingface.co\u002Fnvidia> (filter `PhysicalAI-`)\n> and the curated [Physical AI collection](https:\u002F\u002Fhuggingface.co\u002Fcollections\u002Fnvidia\u002Fphysical-ai).\n> When upstream cards drift, re-check the HF page; this skill mirrors\n> the cards as of Apr 2026.\n\n## Prerequisites\n\n- HuggingFace account with the **dataset card opened in a browser at\n  least once**, and the gating checkbox accepted on every dataset you\n  intend to download.\n- HuggingFace user access token exported as `HF_TOKEN` (create at\n  \u003Chttps:\u002F\u002Fhuggingface.co\u002Fsettings\u002Ftokens>).\n- `git`, `git-lfs`, and `huggingface_hub[cli]` on PATH.\n- Storage room sized to the dataset you're pulling (see the per-row\n  size column; some are \u003C 1 GB, the AV dataset is 133 TB — always\n  pre-filter with `--include` or `physical_ai_av`).\n\n### Verifying secrets safely\n\n**Always check token presence with `hf auth whoami` or a length-only\nshell test; never write ad-hoc bash that interpolates `HF_TOKEN`\nvalues.** The common one-liner\n\n```bash\n# BAD — leaks the secret to the terminal when the variable is set\necho \"HF_TOKEN: ${HF_TOKEN:+yes}${HF_TOKEN:-no}\"\n```\n\nprints `yes\u003Ctoken-value>` whenever `HF_TOKEN` is set, because\n`${VAR:-no}` only falls back to \"no\" when `VAR` is empty — when set\nit expands to `$VAR`. Use one of these instead:\n\n```bash\nhf auth whoami                              # confirms the token without echoing it\ntest -n \"$HF_TOKEN\" && echo \"HF_TOKEN: set (${#HF_TOKEN} chars)\" || echo \"HF_TOKEN: missing\"\n```\n\nRotate any token you suspect was echoed at\n\u003Chttps:\u002F\u002Fhuggingface.co\u002Fsettings\u002Ftokens>.\n\n## Table of Contents\n\n1. [Common download recipe](#common-download-recipe) — HF auth, gating, CLI.\n2. [Filtered AV download recipe](#filtered-av-download-recipe) — default `hyperion_8.1` filter for `PhysicalAI-Autonomous-Vehicles` raw pulls.\n3. [Pick a dataset by task](#pick-a-dataset-by-task) — fast lookup table.\n4. [Autonomous Vehicles](#autonomous-vehicles) — 5 datasets.\n5. [Robotics — Manipulation](#robotics--manipulation) — 6 datasets.\n6. [Robotics — GR00T](#robotics--gr00t) — 7 datasets.\n7. [Robotics — mindmap](#robotics--mindmap) — 4 datasets.\n8. [Robotics — NuRec \u002F Sim-Ready scenes](#robotics--nurec--sim-ready-scenes) — 2 datasets.\n9. [Robotics — Healthcare](#robotics--healthcare) — 1 dataset.\n10. [Robotics — Grasping](#robotics--grasping) — 1 dataset.\n11. [Robotics — Physical \u002F material properties](#robotics--physical--material-properties) — 2 datasets.\n12. [Spatial Intelligence + SimReady scenes](#spatial-intelligence--simready-scenes) — 5 datasets.\n13. [Community \u002F sample](#community--sample) — 1 dataset.\n14. [License decision tree](#license-decision-tree) — what you can do with each.\n15. [Cross-skill usage map](#cross-skill-usage-map) — which skill consumes which dataset.\n\n## Common download recipe\n\nAll NVIDIA Physical AI datasets live on `huggingface.co\u002Fdatasets\u002Fnvidia\u002F...`\nand use the same access shape:\n\n```bash\nsudo apt -y install git git-lfs\ngit lfs install\n\nuv tool install -U \"huggingface_hub[cli]\"   # or: pip install --upgrade \"huggingface_hub[cli]\"\nhf auth login                                # paste user access token\n```\n\nThe token must:\n\n1. Have a HF user account that's **logged in** to the dataset page in a browser AT LEAST ONCE.\n2. Have **accepted** any license \u002F terms-of-use checkbox the dataset shows\n   (re-accept if the dataset has been re-gated — common for AV).\n\nThree download patterns:\n\n```bash\n# Whole dataset (small \u002F medium)\nhf download nvidia\u002F\u003Cdataset> --repo-type dataset --local-dir .\u002F\u003Cdataset>\n\n# Sub-folder only (recommended for large multi-task collections)\nhf download nvidia\u002FPhysicalAI-Robotics-GR00T-X-Embodiment-Sim \\\n  --repo-type dataset \\\n  --include \"gr1_arms_only.CanSort\u002F**\" \\\n  --local-dir .\u002Fgr00t_dataset\n\n# Sparse-checkout via git-LFS (if you want incremental git-style work)\ngit clone --filter=blob:none --no-checkout https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002F\u003Cdataset>\ncd \u003Cdataset>\ngit sparse-checkout init --cone\ngit sparse-checkout set \u003Csubfolder>\ngit checkout main\n```\n\nSpecial-case downloaders:\n\n- **PhysicalAI-Autonomous-Vehicles** (133 TB) — use the official Python\n  toolkit `pip install physical_ai_av`\n  ([NVlabs\u002Fphysical_ai_av](https:\u002F\u002Fgithub.com\u002FNVlabs\u002Fphysical_ai_av))\n  to filter by sensor \u002F country \u002F split before downloading; otherwise\n  you will pull TBs you don't need.\n\n  **Default policy: pre-filter to `platform_class == 'hyperion_8.1'`\n  before any bulk pull.** The dataset mixes two sensor rigs\n  (`hyperion_8` and `hyperion_8.1`) and the entire downstream\n  NuRec \u002F NCore \u002F Asset-Harvester chain in this repo is *only*\n  validated for `hyperion_8.1`. Roughly half of the 306k clips fall\n  outside that platform, so the filter typically cuts the working set\n  (and download size) in half.\n\n  Apply this filter by default in any download script you generate.\n  Only skip the filter when the user **explicitly** asks for it (e.g.\n  \"download all platforms\", \"include hyperion_8\", \"don't filter by\n  platform\") or when they hand you a specific clip UUID — in that\n  single-clip case, `download_clip_features(clip_id=...)` is already\n  scoped and no platform filter is needed.\n\n  Canonical filtered recipe (see § [Filtered AV download recipe](#filtered-av-download-recipe)\n  below for an end-to-end example):\n\n  ```python\n  from physical_ai_av import PhysicalAIAVDatasetInterface\n\n  dataset = PhysicalAIAVDatasetInterface()\n  dataset.download_metadata()\n  dc = dataset.metadata['data_collection']\n  hyperion_81_clip_ids = dc[dc['platform_class'] == 'hyperion_8.1'].index.tolist()\n  ```\n- **PhysicalAI-Autonomous-Vehicle-Cosmos-Drive-Dreams** (3 TB) — use the\n  upstream `download.py`\n  ([nv-tlabs\u002FCosmos-Drive-Dreams](https:\u002F\u002Fgithub.com\u002Fnv-tlabs\u002FCosmos-Drive-Dreams\u002Fblob\u002Fmain\u002Fscripts\u002Fdownload.py))\n  with `--file_types {hdmap,lidar,synthetic}` to select layers.\n- **PhysicalAI-SpatialIntelligence-Lyra-SDG** (25 TB) — `hf download\n  --local-dir lyra_dataset\u002Ftar`; untar each tar yourself.\n- **Spatial-Intelligence-Warehouse** — chunked TAR-GZs need a manual\n  loop after download (script provided in the upstream card).\n\nFor dataset filtering \u002F preview: NVIDIA's\n[Cosmos Dataset Search (CDS)](https:\u002F\u002Fbuild.nvidia.com\u002Fnvidia\u002Fcosmos-dataset-search)\nlets you query a 41K subset of the AV dataset semantically before\ndownloading.\n\n## Filtered AV download recipe\n\nAlways use this recipe (or a derivative of it) when the user asks to\ndownload raw clips from `PhysicalAI-Autonomous-Vehicles`. The\n`hyperion_8.1` filter is the **default**; it matches every downstream\nskill in this repo (`ncore`, `nre`, `asset-harvester`, `nurec-fixer`).\n\nWhen to **skip** the platform filter:\n\n1. The user explicitly opts out — e.g. \"download all platforms\",\n   \"include hyperion_8 too\", \"ignore platform_class\", or asks for a\n   dataset-wide statistic. In that case, drop the platform mask and\n   warn them that NuRec \u002F NCore tooling will not work on the\n   `hyperion_8` clips.\n2. The user gave you a **specific clip UUID**. A single-clip\n   `download_clip_features(clip_id=...)` is already scoped, and\n   filtering by `platform_class` for one clip is pointless.\n\nRecipe (defaults: filter on; user may layer extra masks like country\nor sensor presence on top):\n\n```python\nfrom physical_ai_av import PhysicalAIAVDatasetInterface\n\ndataset = PhysicalAIAVDatasetInterface()\n\ndataset.download_metadata()\ndc = dataset.metadata['data_collection']\n\nclip_mask = dc['platform_class'] == 'hyperion_8.1'\n\n# Optional extra masks (only add when the user asked for them):\n# clip_mask &= dc['country'] == 'US'\n# sp = dataset.metadata['feature_presence']   # 26.03+; was 'sensor_presence' in 25.10\n# clip_mask &= sp['lidar_top_360fov']\n\nclip_ids = dc[clip_mask].index.tolist()\nprint(f\"Downloading {len(clip_ids)} hyperion_8.1 clips\")\n\ndataset.download_clip_features(\n    clip_id=clip_ids,\n    features=[\"camera_front_wide_120fov\", \"lidar_top_360fov\", \"egomotion\"],\n    max_workers=8,\n)\n```\n\nSingle-clip fast path (no platform filter — the UUID is already\nspecific):\n\n```python\ndataset.download_clip_features(\n    clip_id=\"\u003Cpaste-clip-uuid>\",\n    features=[\"camera_front_wide_120fov\", \"lidar_top_360fov\", \"egomotion\"],\n)\n```\n\nExplicit opt-out (only when the user asked for it):\n\n```python\nclip_ids = dc.index.tolist()   # NO platform filter — all 306k clips\n# WARNING: downstream NuRec \u002F NCore \u002F Asset-Harvester only handle\n# the hyperion_8.1 subset.\n```\n\n## Pick a dataset by task\n\n| Goal | Recommended dataset(s) |\n|------|------------------------|\n| End-to-end AV training (real, multi-sensor) | `PhysicalAI-Autonomous-Vehicles` (133 TB, 1700 h, 25 countries) |\n| AV in NCore V4 format (drop-in for [`ncore`](..\u002Fncore\u002FSKILL.md)) | `PhysicalAI-Autonomous-Vehicles-NCore` (~1.1k clips) |\n| AV photoreal Sim2Real \u002F weather augmentation | `PhysicalAI-Autonomous-Vehicle-Cosmos-Drive-Dreams` (3 TB; 7 weather variants) |\n| AV neural reconstructions ready for CARLA \u002F NuRec | `PhysicalAI-Autonomous-Vehicles-NuRec` (918 USDZ scenes) |\n| Asset Harvester \u002F 3DGS extraction sample clip | `PhysicalAI-Autonomous-Vehicles-NCore` |\n| GR00T post-training, broad coverage | `PhysicalAI-Robotics-GR00T-X-Embodiment-Sim` (1.91 TB, 24 GR1 task families + bimanual + RoboCasa) |\n| GR00T fine-tune on industrial tasks | `PhysicalAI-GR00T-Tuned-Tasks` (Nut Pouring, Exhaust Pipe Sorting) |\n| GR00T eval images \u002F videos | `PhysicalAI-Robotics-GR00T-Eval`, `PhysicalAI-Robotics-GR00T-GR1` |\n| Real humanoid teleop (Unitree G1) | `PhysicalAI-Robotics-GR00T-Teleop-G1` (1000 trajectories) |\n| Sim humanoid teleop (Fourier GR1) | `PhysicalAI-Robotics-GR00T-Teleop-Sim` (24 tasks × 1k trajectories) |\n| Massive humanoid pretraining (44k h, DreamDojo) | `PhysicalAI-Robotics-GR00T-Teleop-GR1` (74.3 GB) |\n| Spatial-memory imitation learning (mindmap) | `PhysicalAI-Robotics-mindmap-{Stick-in-Bin,Drill-in-Box,Cube-Stacking,Mug-in-Drawer}` |\n| Robot pick-place in kitchen (bimanual Kinova Gen3) | `PhysicalAI-Robotics-Manipulation-Kitchen`, `-Manipulation-Objects` |\n| Robot pick-place tabletop (single Franka) | `PhysicalAI-Robotics-Manipulation-SingleArm` |\n| Cosmos-Transfer1 visual-augmented stacking | `PhysicalAI-Robotics-Manipulation-Augmented` |\n| Massive teleop in kitchen (Franka + mobile base) | `PhysicalAI-Robotics-Manipulation-Kitchen-Demos` (600 h, 316 tasks, 55k traj) |\n| MJCF kitchen objects + fixtures (MuJoCo) | `PhysicalAI-Robotics-Manipulation-Objects-Kitchen-MJCF` |\n| Sim-Ready warehouse for IsaacSim | `PhysicalAI-SimReady-Warehouse-01` (753 USD assets) |\n| GR1 tabletop digital cousins (assets) | `PhysicalAI-DigitalCousin-Assets` |\n| 3DGS \u002F Sim-Ready indoor scenes for AMR sim | `PhysicalAI-Robotics-NuRec` (Nova Carter labs, Zurich offices, hand-held) |\n| Multi-cam tracking + 3D box benchmark | `PhysicalAI-SmartSpaces` (AI City Challenge 2024 + 2025) |\n| 3D scene QA \u002F VLM training (warehouses) | `PhysicalAI-Spatial-Intelligence-Warehouse` (499k QA pairs) |\n| Generative 3D scene reconstruction training | `PhysicalAI-SpatialIntelligence-Lyra-SDG` (25 TB; GEN3C-derived) |\n| Radiance-field photometric benchmark | `PhysicalAI-NuRec-PPISP` (8 sequences, +\u002F-2 EV bracketing) |\n| Grasping models (Franka, Robotiq-2f-140, suction) | `PhysicalAI-Robotics-GraspGen` (57M grasps, Objaverse-LVIS) |\n| Healthcare \u002F surgical robotics autonomy | `PhysicalAI-Robotics-Open-H-Embodiment` (750 h, 4.5 TB, 30+ orgs) |\n| Volumetric mechanical \u002F material properties | `PhysicalAI-Robotics-PhysicalAssets-VoMP`, `-VoMP-Eval` |\n\n## Autonomous Vehicles\n\n### PhysicalAI-Autonomous-Vehicles\n\nThe flagship real-world AV dataset.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Autonomous-Vehicles> |\n| Size | **133 TB** — 306,152 clips × 20 s = **1700 h** |\n| Sensors | 7 cameras (1080p @ 30 FPS), top-360 LiDAR (10 Hz, 298k clips), up to 10 radars (160k clips), ego motion, calibration, machine labels |\n| Geography | 25 countries, 2500+ cities (US 155k, Germany 44k, …) |\n| Format | Per-sensor parquet\u002Fmp4 chunks of ~100 clips; UUIDs cross-link sensors |\n| License | **NVIDIA AV Dataset License Agreement** (gated; AV-development-only purpose; no biometric \u002F surveillance \u002F re-identification; expires 12 months after download) |\n| Toolkit | `pip install physical_ai_av` — direct filtered downloads + format docs. **Default to filtering `metadata['data_collection']['platform_class'] == 'hyperion_8.1'` before any bulk pull** (see [Filtered AV download recipe](#filtered-av-download-recipe)); only skip the filter on explicit user request or when downloading a specific clip UUID. |\n| Versions | 26.03 (current; offline-optimized features for 97 % of clips), 25.10 (initial) |\n| Subset preview | 41k clips searchable on [Cosmos Dataset Search](https:\u002F\u002Fbuild.nvidia.com\u002Fnvidia\u002Fcosmos-dataset-search) |\n| Use with | [`..\u002Fncore\u002FSKILL.md`](..\u002Fncore\u002FSKILL.md) (convert raw clips to NCore V4), [`..\u002Fasset-harvester\u002FSKILL.md`](..\u002Fasset-harvester\u002FSKILL.md) (extract per-object Gaussian assets). NuRec workflows are only validated for `platform_class == hyperion_8.1`. Upstream sim\u002Ftraining tools without an in-repo skill: `NVlabs\u002Falpamayo-1.5`, `NVlabs\u002Falpasim`, CARLA. |\n\n### PhysicalAI-Autonomous-Vehicles-NCore\n\nCurated NCore V4 subset of the above.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Autonomous-Vehicles-NCore> |\n| Size | ~1.1k clips with accurate offline calibration \u002F egomotion \u002F cuboids |\n| Format | NCore V4 — `pai_\u003Cuuid>.json` + per-sensor `.zarr.itar` files |\n| License | NVIDIA AV Dataset License Agreement (gated, same as above) |\n| Use with | [`..\u002Fncore\u002FSKILL.md`](..\u002Fncore\u002FSKILL.md) (drop-in), [`..\u002Fasset-harvester\u002FSKILL.md`](..\u002Fasset-harvester\u002FSKILL.md) (sample clip path: `clips\u002F2a6f330-5ab0-4e92-99d4-d19e406952f4\u002F`) |\n| Notes | Built via [PAI data converter](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fncore\u002Ftree\u002Fmain\u002Ftools\u002Fdata_converter\u002Fpai). Use this BEFORE the full AV dataset for any NCore-driven workflow. |\n\n### PhysicalAI-Autonomous-Vehicles-NuRec\n\nPre-built NuRec dynamic neural reconstructions ready for IsaacSim \u002F CARLA.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Autonomous-Vehicles-NuRec> |\n| Size | 918 USDZ scenes, ~20 s each, with surface meshes + front-camera mp4 + `labels.json` (Batch0002+) |\n| Reconstruction | 6 cameras (front-wide 120°, front-tele 30°, cross-L\u002FR 120°, rear-L\u002FR 70°) |\n| Versions | 26.02 (current), 25.07, 25.05 |\n| License | NVIDIA AV Dataset License Agreement (gated) |\n| Use with | [`..\u002Fnre\u002FSKILL.md`](..\u002Fnre\u002FSKILL.md) (render the USDZs locally or over `serve-grpc`), [`..\u002Fnurec-fixer\u002FSKILL.md`](..\u002Fnurec-fixer\u002FSKILL.md) (clean up rendered frames). Upstream consumer without an in-repo skill: CARLA (NuRec integration in 0.9.16+). |\n\n### PhysicalAI-Autonomous-Vehicle-Cosmos-Drive-Dreams\n\nCosmos-Transfer-style synthetic + HD-map labels for diverse weather.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Autonomous-Vehicle-Cosmos-Drive-Dreams> |\n| Size | 3 TB total (synthetic only ~700 GB) — 5,843 RDS-HQ clips × 2 chunks × 7 weather = 81,802 synthetic videos (121 frames each) |\n| Modalities | Cosmos-generated MP4, HDMap (lanes\u002Flanelines\u002Froad boundaries\u002Fwait lines\u002Fcrosswalks\u002Fmarkings\u002Fpoles\u002Flights\u002Fsigns), LiDAR, vehicle pose, camera intrinsics (ftheta + pinhole), 4D object tracking |\n| Cameras | 7 (front-wide\u002Fcross-L\u002Fcross-R\u002Frear-L\u002Frear-R\u002Frear-tele\u002Ffront-tele) |\n| Weather variants | Foggy \u002F Golden hour \u002F Morning \u002F Night \u002F Rainy \u002F Snowy \u002F Sunny |\n| License | **CC-BY-4.0** (commercial OK with attribution) |\n| Tooling | `wget … scripts\u002Fdownload.py; python download.py --odir \u003Cpath> --file_types hdmap,lidar,synthetic` |\n| Paper | \u003Chttps:\u002F\u002Farxiv.org\u002Fabs\u002F2506.09042> |\n| Use with | Upstream consumers without an in-repo skill: `nvidia\u002FCosmos-Transfer1`, `nvidia\u002FCosmos-Predict`, `NVlabs\u002Falpasim`, CARLA. |\n\n### PhysicalAI-Autonomous-Vehicle-Cosmos-Synthetic\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Autonomous-Vehicle-Cosmos-Synthetic> |\n| Status | **Pointer \u002F placeholder** — content moved to `PhysicalAI-Autonomous-Vehicle-Cosmos-Drive-Dreams`. Use that. (Card is 2.59 kB.) |\n\n## Robotics — Manipulation\n\nAll in **LeRobot v2.x** format unless noted, generated in IsaacSim with\ntask-and-motion planning + `scene_synthesizer` procedural scenes +\nCuRobo motion generation.\n\n### PhysicalAI-Robotics-Manipulation-Kitchen\n\nBimanual Kinova Gen3 in procedurally-generated kitchens.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-Manipulation-Kitchen> |\n| Size | 12 GB total |\n| Tasks | open\u002Fclose × {cabinet, dishwasher, fridge, drawer} = 8 |\n| Trajectories | ~874 episodes total (range 72–205 per task) |\n| Cameras | 6 × 512² RGB+depth+segmentation (world \u002F external \u002F each wrist \u002F head) |\n| License | CC-BY-4.0 |\n| Commercial | ✅ |\n\n### PhysicalAI-Robotics-Manipulation-Objects\n\nSame kitchen environment, bimanual Kinova; pick \u002F place bench \u002F place cabinet.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-Manipulation-Objects> |\n| Size | 4.26 GB |\n| Tasks | `pick`, `place_bench`, `place_cabinet` (540 episodes total) |\n| License | CC-BY-4.0 (intended R&D only per card) |\n| Use with | Upstream Isaac Sim \u002F Isaac Lab (no in-repo skill). |\n\n### PhysicalAI-Robotics-Manipulation-SingleArm\n\nFranka Panda tabletop, procedurally generated.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-Manipulation-SingleArm> |\n| Size | 15.3 GB |\n| Tasks | `panda-stack-wide`, `panda-stack-platforms`, `panda-stack-platforms-texture`, `panda-open-cabinet-{left,right}`, `panda-open-drawer` (~38k episodes) |\n| Modalities | World cam + wrist cam (RGB + depth on the texture\u002Fcabinet\u002Fdrawer subsets) |\n| State | 53 D (stack-wide) \u002F 81 D (others) — proprioception + object poses |\n| License | CC-BY-4.0; commercial OK |\n\n### PhysicalAI-Robotics-Manipulation-Augmented\n\nMimic-generated Franka cube-stacking, plus Cosmos-Transfer1 visual augmentation.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-Manipulation-Augmented> |\n| Size | 77.9 GB |\n| Episodes | 1000 mimic + 1000 Cosmos-augmented (table + wrist cams, depth + seg + normals) |\n| Trick | 10 human teleops → MimicGen 1k → Cosmos Transfer1 photoreal domain randomization |\n| License | CC-BY-4.0; commercial OK |\n| Paper | \u003Chttps:\u002F\u002Farxiv.org\u002Fabs\u002F2503.14492> (Cosmos-Transfer1) |\n| Use with | Upstream consumers without an in-repo skill: `nvidia\u002FCosmos-Transfer1` (legacy Transfer1 workflow), Isaac Sim \u002F Isaac Lab (replay scripts ship in the dataset repo). |\n\n### PhysicalAI-Robotics-Manipulation-Kitchen-Demos\n\nMassive human-teleop dataset on Franka + Omron mobile base.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-Manipulation-Kitchen-Demos> |\n| Size | 600 h, 55k trajectories, 316 tasks |\n| Format | LeRobot v2.x with MuJoCo `extras\u002F` (`model.xml.gz` + raw states) |\n| Cameras | left + right agentview + eye-in-hand |\n| Tasks | `pretrain\u002Fatomic\u002F...` × 100 traj\u002Ftask (Open*, Close*, PickPlace*, Adjust*, Coffee*, NavigateKitchen, …) |\n| Use with | Pair with the MJCF assets dataset below for replay in MuJoCo. |\n\n### PhysicalAI-Robotics-Manipulation-Objects-Kitchen-MJCF\n\nThe MuJoCo XML assets that back the Kitchen-Demos environment.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-Manipulation-Objects-Kitchen-MJCF> |\n| Size | 1.32 GB |\n| Categories | Objects (~58 categories from kettle to whisk) + Fixtures (12 — blender, coffee machine, dishwasher, electric kettle, fridge, microwave, oven, stand mixer, stove, toaster, toaster oven, cabinet panel) |\n| Format | Per-model `model.xml` + visual \u002F collision OBJ + textures, zipped per category |\n| Use with | MuJoCo replay of `Manipulation-Kitchen-Demos`. |\n\n## Robotics — GR00T\n\nGR00T = NVIDIA's generalist humanoid foundation model line. Most data is\nsim-generated for post-training; eval \u002F real-robot supplements are\nsmall.\n\n### PhysicalAI-Robotics-GR00T-X-Embodiment-Sim\n\nLargest GR00T post-training corpus.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-GR00T-X-Embodiment-Sim> |\n| Size | **1.91 TB** |\n| Composition | 9k cross-embodied bimanual (Panda + GR1) + 240k humanoid GR1 tabletop + 24k downsampled + 72k single-Panda RoboCasa + 102 Unitree G1 loco-manipulation = ~345k trajectories |\n| Used by | `nvidia\u002FGR00T-N1.5-3B`, `GR00T-N1.6-3B`, `GR00T-N1.6-bridge`, `GR00T-N1.6-G1-PnPAppleToPlate`, `GR00T-N1.6-DROID`, `GR00T-N1.6-fractal` |\n| Download tip | Always pass `--include \"\u003Ctask>\u002F**\"` — full clone is 1.91 TB |\n\n### PhysicalAI-GR00T-Tuned-Tasks\n\nTwo industrial post-training task families.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-GR00T-Tuned-Tasks> |\n| Size | 26.5 GB |\n| Tasks | Exhaust-Pipe-Sorting (1000), Nut-Pouring (1000) |\n| Format | HDF5 + GR00T-LeRobot, 256² first-person RGB, 26-DoF state\u002Faction, 20 Hz |\n| License | CC-BY-4.0; commercial OK |\n| Models | `nvidia\u002FGR00T-N1-2B-tuned-Nut-Pouring-task`, `…-Exhaust-Pipe-Sorting-task` |\n\n### PhysicalAI-Robotics-GR00T-Teleop-G1\n\nReal-robot Unitree G1 fruit pick-and-place.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-GR00T-Teleop-G1> |\n| Size | 534 MB |\n| Trajectories | 1000 real teleop, Unitree G1 upper body + Tri-finger hands + RealSense |\n| Tasks | Pick {apple, pear, grapes, starfruit} → basket |\n| Format | MP4 + HDF5 |\n| License | CC-BY-4.0; commercial OK |\n| Use with | `Isaac-GR00T` finetune docs (`getting_started\u002F3_0_new_embodiment_finetuning.md`) |\n\n### PhysicalAI-Robotics-GR00T-Teleop-Sim\n\nSimulated GR1 tabletop teleop.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-GR00T-Teleop-Sim> |\n| Size | 55.4 GB (39 GB LeRobot + 14 GB HDF5) |\n| Trajectories | 24 tasks × 1000 each |\n| License | **CC-BY-NC-4.0** (non-commercial) — different from G1 above |\n| Format | HDF5 + LeRobot |\n\n### PhysicalAI-Robotics-GR00T-Teleop-GR1\n\nDreamDojo pretraining corpus — large-scale human egocentric video.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-GR00T-Teleop-GR1> |\n| Size | 74.3 GB |\n| Coverage | 44k hours of human egocentric data (per project page) |\n| Project | \u003Chttps:\u002F\u002Fdreamdojo-world.github.io\u002F> + \u003Chttps:\u002F\u002Fgithub.com\u002FNVIDIA\u002FDreamDojo> |\n| Paper | \u003Chttps:\u002F\u002Farxiv.org\u002Fabs\u002F2602.06949> |\n\n### PhysicalAI-Robotics-GR00T-GR1\n\nLab-recorded GR1-T2 third-person video.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-GR00T-GR1> |\n| Size | 142 MB |\n| Records | 92 MP4 videos (Fourier GR1-T2) |\n| Use | DreamGen training reference |\n| License | CC-BY-4.0; commercial OK |\n\n### PhysicalAI-Robotics-GR00T-Eval\n\nGR00T eval initial-state frames.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-GR00T-Eval> |\n| Size | 180 MB |\n| Records | 123 PNG frames (GR1-T2 robot's first-person view) + per-frame TXT |\n| License | CC-BY-4.0 |\n\n## Robotics — mindmap\n\nSpatial-memory benchmark from `nvidia-isaac\u002Fnvblox_mindmap`. Each\ndataset is one task with the same multimodal layout (RGB-D + camera\nintrinsics\u002Fposes + nvblox vertex features in `.zst` + robot state).\n**All four are CC-BY-NC-4.0** (research only). Models trained:\n[`nvidia\u002FPhysicalAI-Robotics-mindmap-Checkpoints`](https:\u002F\u002Fhuggingface.co\u002Fnvidia\u002FPhysicalAI-Robotics-mindmap-Checkpoints).\n\n| Dataset | Robot | Teleop tool | Demos provided | Total mimic-generated | Storage |\n|---------|-------|-------------|----------------|------------------------|---------|\n| [`PhysicalAI-Robotics-mindmap-GR1-Stick-in-Bin`](https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-mindmap-GR1-Stick-in-Bin) | Fourier GR1 | Apple Vision Pro | 10 (`mindmap` fmt) + HDF5 | 200 (from 20 human teleops) | 103 GB |\n| [`PhysicalAI-Robotics-mindmap-GR1-Drill-in-Box`](https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-mindmap-GR1-Drill-in-Box) | Fourier GR1 | Apple Vision Pro | 10 + HDF5 | 200 (from 20 human teleops) | 53.7 GB |\n| [`PhysicalAI-Robotics-mindmap-Franka-Cube-Stacking`](https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-mindmap-Franka-Cube-Stacking) | Franka | SpaceMouse | 10 + HDF5 | 1000 (from 10 human teleops) | 6.12 GB |\n| [`PhysicalAI-Robotics-mindmap-Franka-Mug-in-Drawer`](https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-mindmap-Franka-Mug-in-Drawer) | Franka | SpaceMouse | 10 + HDF5 | 250 (from 15 human teleops) | 34.5 GB |\n\n> Provided datasets ship 10 mindmap-formatted demos for storage reasons;\n> regenerate the full 200\u002F1000\u002F250 with the upstream\n> [data-generation docs](https:\u002F\u002Fnvidia-isaac.github.io\u002Fnvblox_mindmap\u002Fpages\u002Fdata_generation.html).\n\nPaper: \u003Chttps:\u002F\u002Farxiv.org\u002Fabs\u002F2509.20297>. Codebase:\n\u003Chttps:\u002F\u002Fgithub.com\u002Fnvidia-isaac\u002Fnvblox_mindmap>.\n\n## Robotics — NuRec \u002F Sim-Ready scenes\n\n### PhysicalAI-Robotics-NuRec\n\nIndoor 3DGUT scenes for IsaacSim AMR simulation.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-NuRec> |\n| Size | 62.9 GB |\n| Scenes | Nova-Carter (galileo, cafe, wormhole) — stereo, with mesh + occupancy; Zurich offices (lounge, fourth-floor iphone) — mono, no mesh; Endeavor hand-held (andoria, livingroom, wormhole) — stereo with mesh |\n| Format | USDZ (3DGUT + mesh + occupancy) loadable in **Isaac Sim 5.1** |\n| Workflows | [Stereo NuRec (Isaac ROS + cuSFM + FoundationStereo + nvblox + 3DGURT)](https:\u002F\u002Fdocs.nvidia.com\u002Fnurec\u002Frobotics\u002Fneural_reconstruction_stereo.html) for Carter; [Mono NuRec (COLMAP + 3DGURT)](https:\u002F\u002Fdocs.nvidia.com\u002Fnurec\u002Frobotics\u002Fneural_reconstruction_mono.html) for Zurich |\n| Gating | Contact-info gate (no separate license) |\n| License | CC-BY-4.0 |\n| Use with | [`..\u002Fnre\u002FSKILL.md`](..\u002Fnre\u002FSKILL.md) to retrain reconstructions; upstream Isaac Sim 5.1 (no in-repo skill) for AMR simulation; pair with [`MobilityGen`](https:\u002F\u002Fgithub.com\u002FNVlabs\u002FMobilityGen) for AMR data generation. |\n\n### PhysicalAI-NuRec-PPISP\n\nPhotometric-variation benchmark for radiance-field methods.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-NuRec-PPISP> |\n| Size | 15.2 GB (8.1 GB COLMAP + 6.2 GB NCore V4) |\n| Captures | 4 outdoor scenes × 3 cameras (Nikon Z7, OM-1 II, iPhone 13 Pro) = 8 sequences (~2600 photos), `+\u002F-2 EV` exposure bracketing |\n| Variants | Standard (full bracket) + auto (re-processed with auto-exposure \u002F WB) |\n| License | CC-BY-4.0; commercial OK |\n| Use with | 3DGRUT \u002F GSplat benchmarking via [`..\u002Fnre\u002FSKILL.md`](..\u002Fnre\u002FSKILL.md) (`eval-rendering-metrics`), [`..\u002Fasset-harvester\u002FSKILL.md`](..\u002Fasset-harvester\u002FSKILL.md). |\n\n## Robotics — Healthcare\n\n### PhysicalAI-Robotics-Open-H-Embodiment\n\nSurgical \u002F ultrasound robotics multi-embodiment corpus.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-Open-H-Embodiment> |\n| Size | 4.5 TB, 750 h, 120,000 trajectories |\n| Format | LeRobot v2.1 — MP4 video + Parquet kinematics + JSONL manifests |\n| Contributors | 30+ orgs (JHU, Stanford, UCSD, UCB, Vanderbilt, TUM, MBZUAI, …) |\n| Purpose | Healthcare autonomy + world-foundation-model training (used by `nvidia\u002FGR00T-H` and `nvidia\u002FCosmos-H-Surgical-Simulator`) |\n| License | CC-BY-4.0 |\n\n## Robotics — Grasping\n\n### PhysicalAI-Robotics-GraspGen\n\nSim2Real grasping at scale.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-GraspGen> |\n| Size | 21.6 GB |\n| Coverage | **57 M+ grasps** over 8515 Objaverse-XL (LVIS) objects |\n| Grippers | Franka Panda, Robotiq-2f-140, suction (30 mm radius) |\n| Format | WebDataset shards (`grasp_data\u002F{franka,robotiq2f140,suction}\u002Fshard_{0-7}.tar`) + train\u002Fvalid splits |\n| License | CC-BY-4.0; commercial OK |\n| Models | [`adithyamurali\u002FGraspGenModels`](https:\u002F\u002Fhuggingface.co\u002Fadithyamurali\u002FGraspGenModels) |\n| Tip | Objaverse meshes are **NOT** included — pull separately via the bundled `download_objaverse.py` |\n\n## Robotics — Physical \u002F material properties\n\nVoMP = Volumetric Mechanical Properties.\n\n### PhysicalAI-Robotics-PhysicalAssets-VoMP\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-PhysicalAssets-VoMP> |\n| Size | 65.9 GB on-disk (full data nominally 125 GB pre-compression) |\n| Records | 1664 objects, 37,337,952 voxels, multi-view renders + VLM material annotations |\n| License | CC-BY-4.0 |\n\n### PhysicalAI-Robotics-PhysicalAssets-VoMP-Eval\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-PhysicalAssets-VoMP-Eval> |\n| Size | 8.41 GB on-disk (full eval data nominally 125 GB pre-compression) |\n| Use | Held-out eval split for the VoMP model |\n| License | CC-BY-4.0 |\n\n## Spatial Intelligence + SimReady scenes\n\n### PhysicalAI-SimReady-Warehouse-01\n\nOpenUSD warehouse scene + asset library.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-SimReady-Warehouse-01> |\n| Size | 14.4 GB, 753 USD assets + master scene (`physical_ai_simready_warehouse_01.usd`) |\n| Asset class | Prop \u002F Assembly \u002F Scenario; 1.1.0 adds physically-graspable subset |\n| Metadata | CSV catalogue with WikiData Q-codes, mass (kg), thumbnails |\n| Target | **Isaac Sim 4.x** (Properties → disable Instanceable → Physics → Rigid Body to make assets dynamic) |\n| License | CC-BY-4.0 |\n| Use with | Upstream consumers without an in-repo skill: Isaac Sim \u002F Isaac Lab, Omniverse SDG. |\n\n### PhysicalAI-DigitalCousin-Assets\n\nCompanion-asset library for the GR1 tabletop sim environments.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-DigitalCousin-Assets> |\n| Size | 270 MB |\n| Content | 3D meshes, textures, metadata for tabletop objects (mug, bottle, bowl, container, …) used by GR1 sim tasks |\n| License | **CC-BY-NC-4.0** (non-commercial) |\n\n### PhysicalAI-SmartSpaces\n\nMulti-camera tracking + 3D box benchmark (AI City Challenge).\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-SmartSpaces> |\n| Size | 3.53 TB total (216 GB MTMC_Tracking_2024 + 3.31 TB MTMC_Tracking_2025) |\n| 2024 | 90 scenes, 212 h, 953 cameras — Person-only, 2D boxes + multi-cam IDs (52M \u002F 135M) |\n| 2025 | 23 scenes, 42 h, 504 cameras — Person\u002FForklift\u002FNovaCarter\u002FTransporter\u002FFourierGR1T2\u002FAgilityDigit, 3D boxes + depth maps (8.9M \u002F 73M) |\n| Splits | Warehouse (train\u002Fval\u002Ftest) + Lab (val) + Hospital (val) + 4 test scenes added 2025-05-28 |\n| Eval | \u003Chttps:\u002F\u002Feval.aicitychallenge.org\u002Faicity2024> + \u003Chttps:\u002F\u002Feval.aicitychallenge.org\u002Faicity2025>; 3D-bbox HOTA metric for 2025 |\n| Paper | \u003Chttps:\u002F\u002Farxiv.org\u002Fabs\u002F2412.00692> (MCBLT) |\n\n### PhysicalAI-Spatial-Intelligence-Warehouse\n\nVLM-style spatial QA in warehouses.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Spatial-Intelligence-Warehouse> |\n| Size | 261 GB |\n| QA pairs | 499k train + 19k test + 1.9k val (categories: `left_right`, `multi_choice_question`, `distance` in metres, `count`) |\n| Imagery | ~95k RGB-D pairs, RLE object masks (pycoco), LLaVA-style conversations |\n| Annotation | Rule-based + Llama-3.1-70B-Instruct refinement |\n| Gating | Contact-info gate (no separate license) |\n| License | CC-BY-4.0 |\n| Format | `train.json` \u002F `val.json` \u002F `test.json` + chunked TAR-GZs of images + depths |\n\n### PhysicalAI-SpatialIntelligence-Lyra-SDG\n\nGEN3C-derived multi-view 3D + 4D training data for `nv-tlabs\u002Flyra`.\n\n| Field | Value |\n|-------|-------|\n| HF | \u003Chttps:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-SpatialIntelligence-Lyra-SDG> |\n| Size | **25 TB** |\n| Composition | 59,031 multi-view 3D examples (354,186 videos) + 7,378 4D examples (44,268 videos), 6 trajectories per source |\n| Modalities | RGB MP4 + camera pose `.npz` + depth zip |\n| License | CC-BY-4.0 |\n| Paper | \u003Chttps:\u002F\u002Farxiv.org\u002Fabs\u002F2509.19296> (Lyra) |\n| Pair with | [`nvidia\u002FLyra-Testing-Example`](https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FLyra-Testing-Example) for inference; [`nv-tlabs\u002Flyra`](https:\u002F\u002Fgithub.com\u002Fnv-tlabs\u002Flyra) for training |\n\n### (See also) PhysicalAI-NuRec-PPISP\n\nListed in [§ 7](#robotics--nurec--sim-ready-scenes) but is also a\nspatial-intelligence radiance-field benchmark.\n\n## License decision tree\n\n| License | Commercial OK? | Reproducible? | Datasets |\n|---------|----------------|----------------|----------|\n| **CC-BY-4.0** | ✅ (with attribution) | ✅ (must keep notice) | most — Cosmos-Drive-Dreams, GraspGen, all Manipulation-* (except where noted), GR00T-Teleop-G1, GR00T-Tuned-Tasks, GR00T-GR1, GR00T-Eval, GR00T-X-Embodiment-Sim, Robotics-NuRec, NuRec-PPISP, Open-H-Embodiment, SmartSpaces (CC-BY-4.0 implied via card), Spatial-Intelligence-Warehouse, SpatialIntelligence-Lyra-SDG, SimReady-Warehouse-01, VoMP \u002F VoMP-Eval, Manipulation-Augmented |\n| **CC-BY-NC-4.0** (non-commercial) | ❌ | ✅ research | GR00T-Teleop-Sim, DigitalCousin-Assets, **all 4 mindmap** datasets |\n| **NVIDIA AV Dataset License Agreement** (gated, AV-only purpose, 12-month expiry) | ✅ ONLY for AV \u002F ADAS development on NVIDIA tech | ❌ — no derivative works, no redistribution, no biometric \u002F re-id \u002F surveillance use | `PhysicalAI-Autonomous-Vehicles`, `…-NCore`, `…-NuRec` |\n\nFor internal NVIDIA use, the auto-derivable rule of thumb:\n\n1. If `Robotics-Manipulation-*` and **not** mindmap \u002F DigitalCousin-Assets → CC-BY-4.0 commercial OK.\n2. If `mindmap-*` → research only (NC).\n3. If `Autonomous-Vehicles*` → AV License only, gated.\n4. Everything else → check the card.\n\n## Cross-skill usage map\n\nSibling skills in this hub are linked by relative path; upstream\nprojects without an in-repo skill are linked by URL.\n\n| Dataset | In-repo sibling skill(s) | Upstream consumers (no in-repo skill) |\n|---------|--------------------------|---------------------------------------|\n| `PhysicalAI-Autonomous-Vehicles` | [`..\u002Fncore\u002FSKILL.md`](..\u002Fncore\u002FSKILL.md) | `NVlabs\u002Falpamayo-1.5`, `NVlabs\u002Falpasim`, CARLA |\n| `…-NCore` | [`..\u002Fncore\u002FSKILL.md`](..\u002Fncore\u002FSKILL.md), [`..\u002Fasset-harvester\u002FSKILL.md`](..\u002Fasset-harvester\u002FSKILL.md) | — |\n| `…-NuRec` | [`..\u002Fnre\u002FSKILL.md`](..\u002Fnre\u002FSKILL.md), [`..\u002Fnurec-fixer\u002FSKILL.md`](..\u002Fnurec-fixer\u002FSKILL.md) | CARLA (NuRec integration 0.9.16+) |\n| `…-Cosmos-Drive-Dreams` | — | `nvidia\u002FCosmos-Transfer1`, `nvidia\u002FCosmos-Predict`, `NVlabs\u002Falpasim`, CARLA |\n| `…-Cosmos-Synthetic` | — | (pointer to Cosmos-Drive-Dreams) |\n| `Robotics-Manipulation-Kitchen` \u002F `-Objects` \u002F `-SingleArm` | — | Isaac Sim \u002F Isaac Lab |\n| `Robotics-Manipulation-Augmented` | — | `nvidia\u002FCosmos-Transfer1` (Transfer1 path), Isaac Sim \u002F Isaac Lab |\n| `Robotics-Manipulation-Kitchen-Demos` + `-Kitchen-MJCF` | — | MuJoCo direct; Isaac Sim for MJCF→USD |\n| `Robotics-GR00T-X-Embodiment-Sim` \u002F `-Tuned-Tasks` | — | [`NVIDIA\u002FIsaac-GR00T`](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FIsaac-GR00T), Isaac Sim |\n| `Robotics-GR00T-Teleop-G1` \u002F `-Sim` \u002F `-GR1` (DreamDojo) | — | [`NVIDIA\u002FIsaac-GR00T`](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FIsaac-GR00T) |\n| `Robotics-GR00T-GR1` (DreamGen ref) \u002F `-Eval` | — | reference assets only |\n| `Robotics-mindmap-*` | — | [`nvidia-isaac\u002Fnvblox_mindmap`](https:\u002F\u002Fgithub.com\u002Fnvidia-isaac\u002Fnvblox_mindmap), Isaac Lab |\n| `Robotics-NuRec` | [`..\u002Fnre\u002FSKILL.md`](..\u002Fnre\u002FSKILL.md) | Isaac Sim 5.1, [`MobilityGen`](https:\u002F\u002Fgithub.com\u002FNVlabs\u002FMobilityGen) |\n| `NuRec-PPISP` | [`..\u002Fnre\u002FSKILL.md`](..\u002Fnre\u002FSKILL.md) (3DGRUT \u002F GSplat benchmarking) | — |\n| `Robotics-Open-H-Embodiment` | — | `nvidia\u002FGR00T-H`, `nvidia\u002FCosmos-H-Surgical-Simulator` |\n| `Robotics-GraspGen` | — | Ships its own visualisation scripts; Isaac Sim for replay |\n| `Robotics-PhysicalAssets-VoMP` \u002F `-Eval` | — | VoMP model |\n| `SimReady-Warehouse-01` | — | Isaac Sim 4.x, Omniverse SDG |\n| `DigitalCousin-Assets` | — | Isaac Sim |\n| `SmartSpaces` | — | AI City Challenge eval server |\n| `Spatial-Intelligence-Warehouse` | — | Warehouse VLM benchmark (no upstream skill) |\n| `SpatialIntelligence-Lyra-SDG` | — | [`nv-tlabs\u002Flyra`](https:\u002F\u002Fgithub.com\u002Fnv-tlabs\u002Flyra), `nvidia\u002FCosmos-Predict` (GEN3C lineage) |\n\n## Verify\n\nAfter downloading any dataset:\n\n```bash\nls \u003Clocal-dir>\ndu -sh \u003Clocal-dir>          # confirm size matches the table above (within 10 %)\n\n# For LeRobot v2.x:\npython -c \"from lerobot.common.datasets.lerobot_dataset import LeRobotDataset; \\\n           d = LeRobotDataset('\u003Clocal-dir>'); print(d.meta.info)\"\n\n# For NCore V4 (.zarr.itar):\nncore_vis \u003Clocal-dir>\u002Fclips\u002F\u003Cuuid>\u002Fpai_\u003Cuuid>.json   # via ..\u002Fncore\u002FSKILL.md\n\n# For USDZ scenes:\n# load in Isaac Sim 5.1 — File → Open → \u003Cscene>.usdz\n```\n\nGREEN when:\n\n- File counts and total size match the dataset card to within ~10 %.\n- For gated datasets, the download did not silently terminate at the\n  license-agreement page (re-`hf auth login` if the first chunk is HTML).\n- For LeRobot datasets, `meta\u002Finfo.json` parses and the episode count\n  matches the card.\n\n## Troubleshooting\n\n- **`Repo gated. Cannot access … 401`** — open the dataset URL in a\n  browser, click *Agree* on the license \u002F contact-info form, **then** retry\n  with the same token. Tokens don't get auto-refreshed when a new\n  agreement appears (re-accept after major version bumps).\n- **First chunk is 5 KB of HTML** — same cause as above.\n- **Download hangs indefinitely on AV (133 TB)** — you almost certainly\n  don't want the whole thing. Use `physical_ai_av` to filter\n  `platform_class == 'hyperion_8.1'` (default — see\n  [Filtered AV download recipe](#filtered-av-download-recipe)), plus\n  any sensor \u002F country \u002F split mask, BEFORE pulling.\n- **AV clip downloaded but downstream NCore \u002F NuRec \u002F Asset-Harvester\n  rejects it** — double-check `data_collection['platform_class']` for\n  that clip; only `hyperion_8.1` is validated. If it's `hyperion_8`,\n  either swap to a `hyperion_8.1` clip or accept that the downstream\n  reconstruction tooling will not work.\n- **Cosmos-Drive-Dreams 3 TB on a small disk** — pass\n  `--file_types synthetic` (700 GB) or `--file_types hdmap` (small) to\n  the official `download.py`.\n- **Lyra-SDG 25 TB out of disk** — `hf download --include \"tar\u002Fstatic_*\"\n  --exclude \"tar\u002Fdynamic_*\"` to take just the 3D half (or vice-versa).\n- **Cosmos-Synthetic looks empty (2.59 kB)** — it's a pointer page\n  redirecting to `Cosmos-Drive-Dreams`. Use that.\n- **mindmap dataset only has 10 demos when the card says 200 \u002F 1000** —\n  expected; regenerate the rest from the HDF5 via the upstream\n  `mindmap` data-generation docs.\n- **AV license expired (12 months)** — re-accept on the HF page; this\n  is by-design per the License Agreement section 7.\n\n## Limitations\n\n- **Catalogue-only.** This skill does not actually run any\n  pipeline — it routes the agent to a sibling skill once the right\n  dataset is chosen.\n- **NVIDIA-only.** Non-NVIDIA datasets (Waymo, nuScenes, KITTI,\n  PandaSet, …) are out of scope; the `ncore` skill covers how to\n  convert *those* into NCore V4 manually.\n- **Gating is enforced by Hugging Face, not by this skill.** Several\n  datasets (e.g. `PhysicalAI-Autonomous-Vehicles*`,\n  `nvidia\u002FFixer`-adjacent assets, mindmap large variants) require\n  *both* license acceptance on the HF page *and* a valid `HF_TOKEN`.\n  Acceptances expire (AV is annual) and must be re-clicked.\n- **Footprint can be enormous.** PhysicalAI-Autonomous-Vehicles is\n  ~133 TB; Lyra-SDG ~25 TB; Cosmos-Drive-Dreams ~3 TB. Always check\n  free disk and use `--include` \u002F `--exclude` filters before kicking\n  off a full pull.\n- **The catalogue drifts.** NVIDIA ships new `PhysicalAI-*` datasets\n  regularly. If a dataset isn't in this skill, browse\n  \u003Chttps:\u002F\u002Fhuggingface.co\u002Fnvidia?search_models=PhysicalAI-> and\n  consider opening a PR to update Section 2 \u002F the relevant family.\n- **No GPU \u002F runtime checks here.** Storage and HF auth are the only\n  prerequisites this skill validates; per-dataset compute needs\n  belong to the consuming sibling skill.\n\n## Source links\n\n- HF org (filter `PhysicalAI-`): \u003Chttps:\u002F\u002Fhuggingface.co\u002Fnvidia>\n- Curated collection (29 items): \u003Chttps:\u002F\u002Fhuggingface.co\u002Fcollections\u002Fnvidia\u002Fphysical-ai>\n- Cosmos Dataset Search (semantic preview of AV): \u003Chttps:\u002F\u002Fbuild.nvidia.com\u002Fnvidia\u002Fcosmos-dataset-search>\n- AV Python toolkit: \u003Chttps:\u002F\u002Fgithub.com\u002FNVlabs\u002Fphysical_ai_av>\n- NCore: \u003Chttps:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fncore>\n- Cosmos-Drive-Dreams downloader: \u003Chttps:\u002F\u002Fgithub.com\u002Fnv-tlabs\u002FCosmos-Drive-Dreams\u002Fblob\u002Fmain\u002Fscripts\u002Fdownload.py>\n- Lyra repo: \u003Chttps:\u002F\u002Fgithub.com\u002Fnv-tlabs\u002Flyra>\n- mindmap repo: \u003Chttps:\u002F\u002Fgithub.com\u002Fnvidia-isaac\u002Fnvblox_mindmap>\n- Isaac-GR00T repo: \u003Chttps:\u002F\u002Fgithub.com\u002FNVIDIA\u002FIsaac-GR00T>\n- DreamDojo project: \u003Chttps:\u002F\u002Fdreamdojo-world.github.io\u002F>\n- AI City Challenge eval: \u003Chttps:\u002F\u002Feval.aicitychallenge.org\u002F>\n",{"data":38,"body":60},{"name":4,"description":6,"version":39,"tools":40,"license":29,"dependencies":44,"compatibility":48,"metadata":49},"0.1.0",[41,42,43],"Shell","Read","Write",[45,46,47],"bash","python","huggingface_hub","All listed datasets are hosted on Hugging Face and require `huggingface_hub` (CLI: `huggingface-cli` \u002F new `hf` shim) plus a HF user access token. Several datasets are GATED — they need explicit license acceptance on the HF dataset page before any token can pull them. Storage ranges from \u003C1 GB to >100 TB (PhysicalAI-Autonomous-Vehicles, 133 TB).",{"author":50,"tags":51,"hf_org":56,"hf_collection":57,"cosmos_dataset_search":58,"pai_av_toolkit":59},"NVIDIA Physical AI",[52,53,54,55,14],"dataset-catalog","physical-ai","huggingface","autonomous-vehicles","https:\u002F\u002Fhuggingface.co\u002Fnvidia","https:\u002F\u002Fhuggingface.co\u002Fcollections\u002Fnvidia\u002Fphysical-ai","https:\u002F\u002Fbuild.nvidia.com\u002Fnvidia\u002Fcosmos-dataset-search","https:\u002F\u002Fgithub.com\u002FNVlabs\u002Fphysical_ai_av",{"type":61,"children":62},"root",[63,72,79,125,143,151,171,177,196,230,236,318,325,350,443,487,610,622,628,809,814,827,955,960,987,992,1337,1342,1567,1580,1585,1635,1647,1689,1694,1873,1878,1914,1919,1950,1955,2506,2510,2515,2520,2765,2770,2775,2922,2927,2932,3065,3070,3075,3245,3251,3312,3317,3337,3342,3347,3457,3463,3468,3572,3577,3582,3713,3718,3723,3844,3849,3854,3972,3977,3982,4082,4087,4092,4097,4102,4239,4244,4249,4356,4361,4366,4486,4491,4496,4585,4590,4595,4697,4702,4707,4792,4797,4802,4873,4878,4918,5129,5145,5165,5170,5175,5180,5344,5349,5354,5484,5489,5494,5499,5609,5614,5619,5624,5780,5785,5790,5795,5866,5872,5943,5948,5953,5958,6080,6085,6090,6167,6172,6177,6308,6313,6318,6489,6494,6506,6654,6660,6672,6677,6809,6814,6865,6870,6875,7563,7569,7574,7809,7814,7848,7854,8059,8065,8207,8213,8335],{"type":64,"tag":65,"props":66,"children":68},"element","h1",{"id":67},"nvidia-physical-ai-datasets-hugging-face",[69],{"type":70,"value":71},"text","NVIDIA Physical AI Datasets (Hugging Face)",{"type":64,"tag":73,"props":74,"children":76},"h2",{"id":75},"purpose",[77],{"type":70,"value":78},"Purpose",{"type":64,"tag":80,"props":81,"children":82},"p",{},[83,85,92,94,100,102,108,110,116,117,123],{"type":70,"value":84},"Help the agent find, evaluate, and download an NVIDIA ",{"type":64,"tag":86,"props":87,"children":89},"code",{"className":88},[],[90],{"type":70,"value":91},"PhysicalAI-*",{"type":70,"value":93},"\ndataset on Hugging Face that fits the user's downstream task —\nautonomous-vehicle reconstruction, robotics manipulation, GR00T\npost-training, spatial-intelligence research, grasping, or sim-ready\ncontent — and then hand off to the sibling skill (",{"type":64,"tag":86,"props":95,"children":97},{"className":96},[],[98],{"type":70,"value":99},"ncore",{"type":70,"value":101},", ",{"type":64,"tag":86,"props":103,"children":105},{"className":104},[],[106],{"type":70,"value":107},"nre",{"type":70,"value":109},",\n",{"type":64,"tag":86,"props":111,"children":113},{"className":112},[],[114],{"type":70,"value":115},"asset-harvester",{"type":70,"value":101},{"type":64,"tag":86,"props":118,"children":120},{"className":119},[],[121],{"type":70,"value":122},"nurec-fixer",{"type":70,"value":124},") or upstream NVIDIA tool that\nactually consumes it.",{"type":64,"tag":80,"props":126,"children":127},{},[128,134,136,141],{"type":64,"tag":129,"props":130,"children":131},"strong",{},[132],{"type":70,"value":133},"Use this skill when:",{"type":70,"value":135}," the user asks \"is there an NVIDIA dataset\nfor X?\", \"where do I get NCore \u002F NuRec \u002F GR00T sample data?\", or is\nshopping the Hugging Face NVIDIA org for ",{"type":64,"tag":86,"props":137,"children":139},{"className":138},[],[140],{"type":70,"value":91},{"type":70,"value":142}," collections.",{"type":64,"tag":80,"props":144,"children":145},{},[146],{"type":64,"tag":129,"props":147,"children":148},{},[149],{"type":70,"value":150},"Do NOT use this skill when:",{"type":64,"tag":152,"props":153,"children":154},"ul",{},[155,161,166],{"type":64,"tag":156,"props":157,"children":158},"li",{},[159],{"type":70,"value":160},"The user already knows the dataset and just wants to run a\npipeline — jump straight to the consuming skill.",{"type":64,"tag":156,"props":162,"children":163},{},[164],{"type":70,"value":165},"The user needs a non-NVIDIA dataset (Waymo, nuScenes, KITTI, …) —\nthis catalog is NVIDIA-only.",{"type":64,"tag":156,"props":167,"children":168},{},[169],{"type":70,"value":170},"The user wants to train Cosmos \u002F GR00T \u002F Isaac Sim itself — that's\nthe upstream tool's job, not this catalog's.",{"type":64,"tag":73,"props":172,"children":174},{"id":173},"overview",[175],{"type":70,"value":176},"Overview",{"type":64,"tag":80,"props":178,"children":179},{},[180,182,187,189,194],{"type":70,"value":181},"Catalog of NVIDIA's open Physical AI dataset family on Hugging Face.\nPick by ",{"type":64,"tag":129,"props":183,"children":184},{},[185],{"type":70,"value":186},"task",{"type":70,"value":188}," (Section 2 § lookup table) or ",{"type":64,"tag":129,"props":190,"children":191},{},[192],{"type":70,"value":193},"family",{"type":70,"value":195}," (Sections 3–10).\nEvery entry lists: dataset path, size, format, license, gating, and the\ndownstream skill in this repo that consumes it.",{"type":64,"tag":197,"props":198,"children":199},"blockquote",{},[200],{"type":64,"tag":80,"props":201,"children":202},{},[203,205,212,214,220,222,228],{"type":70,"value":204},"Source of truth: ",{"type":64,"tag":206,"props":207,"children":210},"a",{"href":56,"rel":208},[209],"nofollow",[211],{"type":70,"value":56},{"type":70,"value":213}," (filter ",{"type":64,"tag":86,"props":215,"children":217},{"className":216},[],[218],{"type":70,"value":219},"PhysicalAI-",{"type":70,"value":221},")\nand the curated ",{"type":64,"tag":206,"props":223,"children":225},{"href":57,"rel":224},[209],[226],{"type":70,"value":227},"Physical AI collection",{"type":70,"value":229},".\nWhen upstream cards drift, re-check the HF page; this skill mirrors\nthe cards as of Apr 2026.",{"type":64,"tag":73,"props":231,"children":233},{"id":232},"prerequisites",[234],{"type":70,"value":235},"Prerequisites",{"type":64,"tag":152,"props":237,"children":238},{},[239,251,272,298],{"type":64,"tag":156,"props":240,"children":241},{},[242,244,249],{"type":70,"value":243},"HuggingFace account with the ",{"type":64,"tag":129,"props":245,"children":246},{},[247],{"type":70,"value":248},"dataset card opened in a browser at\nleast once",{"type":70,"value":250},", and the gating checkbox accepted on every dataset you\nintend to download.",{"type":64,"tag":156,"props":252,"children":253},{},[254,256,262,264,270],{"type":70,"value":255},"HuggingFace user access token exported as ",{"type":64,"tag":86,"props":257,"children":259},{"className":258},[],[260],{"type":70,"value":261},"HF_TOKEN",{"type":70,"value":263}," (create at\n",{"type":64,"tag":206,"props":265,"children":268},{"href":266,"rel":267},"https:\u002F\u002Fhuggingface.co\u002Fsettings\u002Ftokens",[209],[269],{"type":70,"value":266},{"type":70,"value":271},").",{"type":64,"tag":156,"props":273,"children":274},{},[275,281,282,288,290,296],{"type":64,"tag":86,"props":276,"children":278},{"className":277},[],[279],{"type":70,"value":280},"git",{"type":70,"value":101},{"type":64,"tag":86,"props":283,"children":285},{"className":284},[],[286],{"type":70,"value":287},"git-lfs",{"type":70,"value":289},", and ",{"type":64,"tag":86,"props":291,"children":293},{"className":292},[],[294],{"type":70,"value":295},"huggingface_hub[cli]",{"type":70,"value":297}," on PATH.",{"type":64,"tag":156,"props":299,"children":300},{},[301,303,309,311,317],{"type":70,"value":302},"Storage room sized to the dataset you're pulling (see the per-row\nsize column; some are \u003C 1 GB, the AV dataset is 133 TB — always\npre-filter with ",{"type":64,"tag":86,"props":304,"children":306},{"className":305},[],[307],{"type":70,"value":308},"--include",{"type":70,"value":310}," or ",{"type":64,"tag":86,"props":312,"children":314},{"className":313},[],[315],{"type":70,"value":316},"physical_ai_av",{"type":70,"value":271},{"type":64,"tag":319,"props":320,"children":322},"h3",{"id":321},"verifying-secrets-safely",[323],{"type":70,"value":324},"Verifying secrets safely",{"type":64,"tag":80,"props":326,"children":327},{},[328,348],{"type":64,"tag":129,"props":329,"children":330},{},[331,333,339,341,346],{"type":70,"value":332},"Always check token presence with ",{"type":64,"tag":86,"props":334,"children":336},{"className":335},[],[337],{"type":70,"value":338},"hf auth whoami",{"type":70,"value":340}," or a length-only\nshell test; never write ad-hoc bash that interpolates ",{"type":64,"tag":86,"props":342,"children":344},{"className":343},[],[345],{"type":70,"value":261},{"type":70,"value":347},"\nvalues.",{"type":70,"value":349}," The common one-liner",{"type":64,"tag":351,"props":352,"children":356},"pre",{"className":353,"code":354,"language":45,"meta":355,"style":355},"language-bash shiki shiki-themes material-theme-lighter material-theme material-theme-palenight","# BAD — leaks the secret to the terminal when the variable is set\necho \"HF_TOKEN: ${HF_TOKEN:+yes}${HF_TOKEN:-no}\"\n","",[357],{"type":64,"tag":86,"props":358,"children":359},{"__ignoreMap":355},[360,372],{"type":64,"tag":361,"props":362,"children":365},"span",{"class":363,"line":364},"line",1,[366],{"type":64,"tag":361,"props":367,"children":369},{"style":368},"--shiki-light:#90A4AE;--shiki-light-font-style:italic;--shiki-default:#546E7A;--shiki-default-font-style:italic;--shiki-dark:#676E95;--shiki-dark-font-style:italic",[370],{"type":70,"value":371},"# BAD — leaks the secret to the terminal when the variable is set\n",{"type":64,"tag":361,"props":373,"children":375},{"class":363,"line":374},2,[376,382,388,394,399,404,409,414,419,424,428,433,438],{"type":64,"tag":361,"props":377,"children":379},{"style":378},"--shiki-light:#6182B8;--shiki-default:#82AAFF;--shiki-dark:#82AAFF",[380],{"type":70,"value":381},"echo",{"type":64,"tag":361,"props":383,"children":385},{"style":384},"--shiki-light:#39ADB5;--shiki-default:#89DDFF;--shiki-dark:#89DDFF",[386],{"type":70,"value":387}," \"",{"type":64,"tag":361,"props":389,"children":391},{"style":390},"--shiki-light:#91B859;--shiki-default:#C3E88D;--shiki-dark:#C3E88D",[392],{"type":70,"value":393},"HF_TOKEN: ",{"type":64,"tag":361,"props":395,"children":396},{"style":384},[397],{"type":70,"value":398},"${",{"type":64,"tag":361,"props":400,"children":402},{"style":401},"--shiki-light:#90A4AE;--shiki-default:#EEFFFF;--shiki-dark:#BABED8",[403],{"type":70,"value":261},{"type":64,"tag":361,"props":405,"children":406},{"style":384},[407],{"type":70,"value":408},":",{"type":64,"tag":361,"props":410,"children":411},{"style":390},[412],{"type":70,"value":413},"+",{"type":64,"tag":361,"props":415,"children":416},{"style":401},[417],{"type":70,"value":418},"yes",{"type":64,"tag":361,"props":420,"children":421},{"style":384},[422],{"type":70,"value":423},"}${",{"type":64,"tag":361,"props":425,"children":426},{"style":401},[427],{"type":70,"value":261},{"type":64,"tag":361,"props":429,"children":430},{"style":384},[431],{"type":70,"value":432},":-",{"type":64,"tag":361,"props":434,"children":435},{"style":401},[436],{"type":70,"value":437},"no",{"type":64,"tag":361,"props":439,"children":440},{"style":384},[441],{"type":70,"value":442},"}\"\n",{"type":64,"tag":80,"props":444,"children":445},{},[446,448,454,456,461,463,469,471,477,479,485],{"type":70,"value":447},"prints ",{"type":64,"tag":86,"props":449,"children":451},{"className":450},[],[452],{"type":70,"value":453},"yes\u003Ctoken-value>",{"type":70,"value":455}," whenever ",{"type":64,"tag":86,"props":457,"children":459},{"className":458},[],[460],{"type":70,"value":261},{"type":70,"value":462}," is set, because\n",{"type":64,"tag":86,"props":464,"children":466},{"className":465},[],[467],{"type":70,"value":468},"${VAR:-no}",{"type":70,"value":470}," only falls back to \"no\" when ",{"type":64,"tag":86,"props":472,"children":474},{"className":473},[],[475],{"type":70,"value":476},"VAR",{"type":70,"value":478}," is empty — when set\nit expands to ",{"type":64,"tag":86,"props":480,"children":482},{"className":481},[],[483],{"type":70,"value":484},"$VAR",{"type":70,"value":486},". Use one of these instead:",{"type":64,"tag":351,"props":488,"children":490},{"className":353,"code":489,"language":45,"meta":355,"style":355},"hf auth whoami                              # confirms the token without echoing it\ntest -n \"$HF_TOKEN\" && echo \"HF_TOKEN: set (${#HF_TOKEN} chars)\" || echo \"HF_TOKEN: missing\"\n",[491],{"type":64,"tag":86,"props":492,"children":493},{"__ignoreMap":355},[494,518],{"type":64,"tag":361,"props":495,"children":496},{"class":363,"line":364},[497,503,508,513],{"type":64,"tag":361,"props":498,"children":500},{"style":499},"--shiki-light:#E2931D;--shiki-default:#FFCB6B;--shiki-dark:#FFCB6B",[501],{"type":70,"value":502},"hf",{"type":64,"tag":361,"props":504,"children":505},{"style":390},[506],{"type":70,"value":507}," auth",{"type":64,"tag":361,"props":509,"children":510},{"style":390},[511],{"type":70,"value":512}," whoami",{"type":64,"tag":361,"props":514,"children":515},{"style":368},[516],{"type":70,"value":517},"                              # confirms the token without echoing it\n",{"type":64,"tag":361,"props":519,"children":520},{"class":363,"line":374},[521,526,531,535,540,545,550,555,559,564,569,573,578,583,587,592,596,600,605],{"type":64,"tag":361,"props":522,"children":523},{"style":378},[524],{"type":70,"value":525},"test",{"type":64,"tag":361,"props":527,"children":528},{"style":390},[529],{"type":70,"value":530}," -n",{"type":64,"tag":361,"props":532,"children":533},{"style":384},[534],{"type":70,"value":387},{"type":64,"tag":361,"props":536,"children":537},{"style":401},[538],{"type":70,"value":539},"$HF_TOKEN",{"type":64,"tag":361,"props":541,"children":542},{"style":384},[543],{"type":70,"value":544},"\"",{"type":64,"tag":361,"props":546,"children":547},{"style":384},[548],{"type":70,"value":549}," &&",{"type":64,"tag":361,"props":551,"children":552},{"style":378},[553],{"type":70,"value":554}," echo",{"type":64,"tag":361,"props":556,"children":557},{"style":384},[558],{"type":70,"value":387},{"type":64,"tag":361,"props":560,"children":561},{"style":390},[562],{"type":70,"value":563},"HF_TOKEN: set (",{"type":64,"tag":361,"props":565,"children":566},{"style":384},[567],{"type":70,"value":568},"${#",{"type":64,"tag":361,"props":570,"children":571},{"style":401},[572],{"type":70,"value":261},{"type":64,"tag":361,"props":574,"children":575},{"style":384},[576],{"type":70,"value":577},"}",{"type":64,"tag":361,"props":579,"children":580},{"style":390},[581],{"type":70,"value":582}," chars)",{"type":64,"tag":361,"props":584,"children":585},{"style":384},[586],{"type":70,"value":544},{"type":64,"tag":361,"props":588,"children":589},{"style":384},[590],{"type":70,"value":591}," ||",{"type":64,"tag":361,"props":593,"children":594},{"style":378},[595],{"type":70,"value":554},{"type":64,"tag":361,"props":597,"children":598},{"style":384},[599],{"type":70,"value":387},{"type":64,"tag":361,"props":601,"children":602},{"style":390},[603],{"type":70,"value":604},"HF_TOKEN: missing",{"type":64,"tag":361,"props":606,"children":607},{"style":384},[608],{"type":70,"value":609},"\"\n",{"type":64,"tag":80,"props":611,"children":612},{},[613,615,620],{"type":70,"value":614},"Rotate any token you suspect was echoed at\n",{"type":64,"tag":206,"props":616,"children":618},{"href":266,"rel":617},[209],[619],{"type":70,"value":266},{"type":70,"value":621},".",{"type":64,"tag":73,"props":623,"children":625},{"id":624},"table-of-contents",[626],{"type":70,"value":627},"Table of Contents",{"type":64,"tag":629,"props":630,"children":631},"ol",{},[632,643,670,681,692,703,714,725,736,747,757,767,777,787,798],{"type":64,"tag":156,"props":633,"children":634},{},[635,641],{"type":64,"tag":206,"props":636,"children":638},{"href":637},"#common-download-recipe",[639],{"type":70,"value":640},"Common download recipe",{"type":70,"value":642}," — HF auth, gating, CLI.",{"type":64,"tag":156,"props":644,"children":645},{},[646,652,654,660,662,668],{"type":64,"tag":206,"props":647,"children":649},{"href":648},"#filtered-av-download-recipe",[650],{"type":70,"value":651},"Filtered AV download recipe",{"type":70,"value":653}," — default ",{"type":64,"tag":86,"props":655,"children":657},{"className":656},[],[658],{"type":70,"value":659},"hyperion_8.1",{"type":70,"value":661}," filter for ",{"type":64,"tag":86,"props":663,"children":665},{"className":664},[],[666],{"type":70,"value":667},"PhysicalAI-Autonomous-Vehicles",{"type":70,"value":669}," raw pulls.",{"type":64,"tag":156,"props":671,"children":672},{},[673,679],{"type":64,"tag":206,"props":674,"children":676},{"href":675},"#pick-a-dataset-by-task",[677],{"type":70,"value":678},"Pick a dataset by task",{"type":70,"value":680}," — fast lookup table.",{"type":64,"tag":156,"props":682,"children":683},{},[684,690],{"type":64,"tag":206,"props":685,"children":687},{"href":686},"#autonomous-vehicles",[688],{"type":70,"value":689},"Autonomous Vehicles",{"type":70,"value":691}," — 5 datasets.",{"type":64,"tag":156,"props":693,"children":694},{},[695,701],{"type":64,"tag":206,"props":696,"children":698},{"href":697},"#robotics--manipulation",[699],{"type":70,"value":700},"Robotics — Manipulation",{"type":70,"value":702}," — 6 datasets.",{"type":64,"tag":156,"props":704,"children":705},{},[706,712],{"type":64,"tag":206,"props":707,"children":709},{"href":708},"#robotics--gr00t",[710],{"type":70,"value":711},"Robotics — GR00T",{"type":70,"value":713}," — 7 datasets.",{"type":64,"tag":156,"props":715,"children":716},{},[717,723],{"type":64,"tag":206,"props":718,"children":720},{"href":719},"#robotics--mindmap",[721],{"type":70,"value":722},"Robotics — mindmap",{"type":70,"value":724}," — 4 datasets.",{"type":64,"tag":156,"props":726,"children":727},{},[728,734],{"type":64,"tag":206,"props":729,"children":731},{"href":730},"#robotics--nurec--sim-ready-scenes",[732],{"type":70,"value":733},"Robotics — NuRec \u002F Sim-Ready scenes",{"type":70,"value":735}," — 2 datasets.",{"type":64,"tag":156,"props":737,"children":738},{},[739,745],{"type":64,"tag":206,"props":740,"children":742},{"href":741},"#robotics--healthcare",[743],{"type":70,"value":744},"Robotics — Healthcare",{"type":70,"value":746}," — 1 dataset.",{"type":64,"tag":156,"props":748,"children":749},{},[750,756],{"type":64,"tag":206,"props":751,"children":753},{"href":752},"#robotics--grasping",[754],{"type":70,"value":755},"Robotics — Grasping",{"type":70,"value":746},{"type":64,"tag":156,"props":758,"children":759},{},[760,766],{"type":64,"tag":206,"props":761,"children":763},{"href":762},"#robotics--physical--material-properties",[764],{"type":70,"value":765},"Robotics — Physical \u002F material properties",{"type":70,"value":735},{"type":64,"tag":156,"props":768,"children":769},{},[770,776],{"type":64,"tag":206,"props":771,"children":773},{"href":772},"#spatial-intelligence--simready-scenes",[774],{"type":70,"value":775},"Spatial Intelligence + SimReady scenes",{"type":70,"value":691},{"type":64,"tag":156,"props":778,"children":779},{},[780,786],{"type":64,"tag":206,"props":781,"children":783},{"href":782},"#community--sample",[784],{"type":70,"value":785},"Community \u002F sample",{"type":70,"value":746},{"type":64,"tag":156,"props":788,"children":789},{},[790,796],{"type":64,"tag":206,"props":791,"children":793},{"href":792},"#license-decision-tree",[794],{"type":70,"value":795},"License decision tree",{"type":70,"value":797}," — what you can do with each.",{"type":64,"tag":156,"props":799,"children":800},{},[801,807],{"type":64,"tag":206,"props":802,"children":804},{"href":803},"#cross-skill-usage-map",[805],{"type":70,"value":806},"Cross-skill usage map",{"type":70,"value":808}," — which skill consumes which dataset.",{"type":64,"tag":73,"props":810,"children":812},{"id":811},"common-download-recipe",[813],{"type":70,"value":640},{"type":64,"tag":80,"props":815,"children":816},{},[817,819,825],{"type":70,"value":818},"All NVIDIA Physical AI datasets live on ",{"type":64,"tag":86,"props":820,"children":822},{"className":821},[],[823],{"type":70,"value":824},"huggingface.co\u002Fdatasets\u002Fnvidia\u002F...",{"type":70,"value":826},"\nand use the same access shape:",{"type":64,"tag":351,"props":828,"children":830},{"className":353,"code":829,"language":45,"meta":355,"style":355},"sudo apt -y install git git-lfs\ngit lfs install\n\nuv tool install -U \"huggingface_hub[cli]\"   # or: pip install --upgrade \"huggingface_hub[cli]\"\nhf auth login                                # paste user access token\n",[831],{"type":64,"tag":86,"props":832,"children":833},{"__ignoreMap":355},[834,867,884,894,933],{"type":64,"tag":361,"props":835,"children":836},{"class":363,"line":364},[837,842,847,852,857,862],{"type":64,"tag":361,"props":838,"children":839},{"style":499},[840],{"type":70,"value":841},"sudo",{"type":64,"tag":361,"props":843,"children":844},{"style":390},[845],{"type":70,"value":846}," apt",{"type":64,"tag":361,"props":848,"children":849},{"style":390},[850],{"type":70,"value":851}," -y",{"type":64,"tag":361,"props":853,"children":854},{"style":390},[855],{"type":70,"value":856}," install",{"type":64,"tag":361,"props":858,"children":859},{"style":390},[860],{"type":70,"value":861}," git",{"type":64,"tag":361,"props":863,"children":864},{"style":390},[865],{"type":70,"value":866}," git-lfs\n",{"type":64,"tag":361,"props":868,"children":869},{"class":363,"line":374},[870,874,879],{"type":64,"tag":361,"props":871,"children":872},{"style":499},[873],{"type":70,"value":280},{"type":64,"tag":361,"props":875,"children":876},{"style":390},[877],{"type":70,"value":878}," lfs",{"type":64,"tag":361,"props":880,"children":881},{"style":390},[882],{"type":70,"value":883}," install\n",{"type":64,"tag":361,"props":885,"children":887},{"class":363,"line":886},3,[888],{"type":64,"tag":361,"props":889,"children":891},{"emptyLinePlaceholder":890},true,[892],{"type":70,"value":893},"\n",{"type":64,"tag":361,"props":895,"children":896},{"class":363,"line":30},[897,902,907,911,916,920,924,928],{"type":64,"tag":361,"props":898,"children":899},{"style":499},[900],{"type":70,"value":901},"uv",{"type":64,"tag":361,"props":903,"children":904},{"style":390},[905],{"type":70,"value":906}," tool",{"type":64,"tag":361,"props":908,"children":909},{"style":390},[910],{"type":70,"value":856},{"type":64,"tag":361,"props":912,"children":913},{"style":390},[914],{"type":70,"value":915}," -U",{"type":64,"tag":361,"props":917,"children":918},{"style":384},[919],{"type":70,"value":387},{"type":64,"tag":361,"props":921,"children":922},{"style":390},[923],{"type":70,"value":295},{"type":64,"tag":361,"props":925,"children":926},{"style":384},[927],{"type":70,"value":544},{"type":64,"tag":361,"props":929,"children":930},{"style":368},[931],{"type":70,"value":932},"   # or: pip install --upgrade \"huggingface_hub[cli]\"\n",{"type":64,"tag":361,"props":934,"children":936},{"class":363,"line":935},5,[937,941,945,950],{"type":64,"tag":361,"props":938,"children":939},{"style":499},[940],{"type":70,"value":502},{"type":64,"tag":361,"props":942,"children":943},{"style":390},[944],{"type":70,"value":507},{"type":64,"tag":361,"props":946,"children":947},{"style":390},[948],{"type":70,"value":949}," login",{"type":64,"tag":361,"props":951,"children":952},{"style":368},[953],{"type":70,"value":954},"                                # paste user access token\n",{"type":64,"tag":80,"props":956,"children":957},{},[958],{"type":70,"value":959},"The token must:",{"type":64,"tag":629,"props":961,"children":962},{},[963,975],{"type":64,"tag":156,"props":964,"children":965},{},[966,968,973],{"type":70,"value":967},"Have a HF user account that's ",{"type":64,"tag":129,"props":969,"children":970},{},[971],{"type":70,"value":972},"logged in",{"type":70,"value":974}," to the dataset page in a browser AT LEAST ONCE.",{"type":64,"tag":156,"props":976,"children":977},{},[978,980,985],{"type":70,"value":979},"Have ",{"type":64,"tag":129,"props":981,"children":982},{},[983],{"type":70,"value":984},"accepted",{"type":70,"value":986}," any license \u002F terms-of-use checkbox the dataset shows\n(re-accept if the dataset has been re-gated — common for AV).",{"type":64,"tag":80,"props":988,"children":989},{},[990],{"type":70,"value":991},"Three download patterns:",{"type":64,"tag":351,"props":993,"children":995},{"className":353,"code":994,"language":45,"meta":355,"style":355},"# Whole dataset (small \u002F medium)\nhf download nvidia\u002F\u003Cdataset> --repo-type dataset --local-dir .\u002F\u003Cdataset>\n\n# Sub-folder only (recommended for large multi-task collections)\nhf download nvidia\u002FPhysicalAI-Robotics-GR00T-X-Embodiment-Sim \\\n  --repo-type dataset \\\n  --include \"gr1_arms_only.CanSort\u002F**\" \\\n  --local-dir .\u002Fgr00t_dataset\n\n# Sparse-checkout via git-LFS (if you want incremental git-style work)\ngit clone --filter=blob:none --no-checkout https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002F\u003Cdataset>\ncd \u003Cdataset>\ngit sparse-checkout init --cone\ngit sparse-checkout set \u003Csubfolder>\ngit checkout main\n",[996],{"type":64,"tag":86,"props":997,"children":998},{"__ignoreMap":355},[999,1007,1081,1088,1096,1117,1134,1160,1174,1182,1191,1235,1261,1284,1319],{"type":64,"tag":361,"props":1000,"children":1001},{"class":363,"line":364},[1002],{"type":64,"tag":361,"props":1003,"children":1004},{"style":368},[1005],{"type":70,"value":1006},"# Whole dataset (small \u002F medium)\n",{"type":64,"tag":361,"props":1008,"children":1009},{"class":363,"line":374},[1010,1014,1019,1024,1029,1034,1039,1044,1049,1054,1059,1064,1068,1072,1076],{"type":64,"tag":361,"props":1011,"children":1012},{"style":499},[1013],{"type":70,"value":502},{"type":64,"tag":361,"props":1015,"children":1016},{"style":390},[1017],{"type":70,"value":1018}," download",{"type":64,"tag":361,"props":1020,"children":1021},{"style":390},[1022],{"type":70,"value":1023}," nvidia\u002F",{"type":64,"tag":361,"props":1025,"children":1026},{"style":384},[1027],{"type":70,"value":1028},"\u003C",{"type":64,"tag":361,"props":1030,"children":1031},{"style":390},[1032],{"type":70,"value":1033},"datase",{"type":64,"tag":361,"props":1035,"children":1036},{"style":401},[1037],{"type":70,"value":1038},"t",{"type":64,"tag":361,"props":1040,"children":1041},{"style":384},[1042],{"type":70,"value":1043},">",{"type":64,"tag":361,"props":1045,"children":1046},{"style":390},[1047],{"type":70,"value":1048}," --repo-type",{"type":64,"tag":361,"props":1050,"children":1051},{"style":390},[1052],{"type":70,"value":1053}," dataset",{"type":64,"tag":361,"props":1055,"children":1056},{"style":390},[1057],{"type":70,"value":1058}," --local-dir",{"type":64,"tag":361,"props":1060,"children":1061},{"style":390},[1062],{"type":70,"value":1063}," .\u002F",{"type":64,"tag":361,"props":1065,"children":1066},{"style":384},[1067],{"type":70,"value":1028},{"type":64,"tag":361,"props":1069,"children":1070},{"style":390},[1071],{"type":70,"value":1033},{"type":64,"tag":361,"props":1073,"children":1074},{"style":401},[1075],{"type":70,"value":1038},{"type":64,"tag":361,"props":1077,"children":1078},{"style":384},[1079],{"type":70,"value":1080},">\n",{"type":64,"tag":361,"props":1082,"children":1083},{"class":363,"line":886},[1084],{"type":64,"tag":361,"props":1085,"children":1086},{"emptyLinePlaceholder":890},[1087],{"type":70,"value":893},{"type":64,"tag":361,"props":1089,"children":1090},{"class":363,"line":30},[1091],{"type":64,"tag":361,"props":1092,"children":1093},{"style":368},[1094],{"type":70,"value":1095},"# Sub-folder only (recommended for large multi-task collections)\n",{"type":64,"tag":361,"props":1097,"children":1098},{"class":363,"line":935},[1099,1103,1107,1112],{"type":64,"tag":361,"props":1100,"children":1101},{"style":499},[1102],{"type":70,"value":502},{"type":64,"tag":361,"props":1104,"children":1105},{"style":390},[1106],{"type":70,"value":1018},{"type":64,"tag":361,"props":1108,"children":1109},{"style":390},[1110],{"type":70,"value":1111}," nvidia\u002FPhysicalAI-Robotics-GR00T-X-Embodiment-Sim",{"type":64,"tag":361,"props":1113,"children":1114},{"style":401},[1115],{"type":70,"value":1116}," \\\n",{"type":64,"tag":361,"props":1118,"children":1120},{"class":363,"line":1119},6,[1121,1126,1130],{"type":64,"tag":361,"props":1122,"children":1123},{"style":390},[1124],{"type":70,"value":1125},"  --repo-type",{"type":64,"tag":361,"props":1127,"children":1128},{"style":390},[1129],{"type":70,"value":1053},{"type":64,"tag":361,"props":1131,"children":1132},{"style":401},[1133],{"type":70,"value":1116},{"type":64,"tag":361,"props":1135,"children":1137},{"class":363,"line":1136},7,[1138,1143,1147,1152,1156],{"type":64,"tag":361,"props":1139,"children":1140},{"style":390},[1141],{"type":70,"value":1142},"  --include",{"type":64,"tag":361,"props":1144,"children":1145},{"style":384},[1146],{"type":70,"value":387},{"type":64,"tag":361,"props":1148,"children":1149},{"style":390},[1150],{"type":70,"value":1151},"gr1_arms_only.CanSort\u002F**",{"type":64,"tag":361,"props":1153,"children":1154},{"style":384},[1155],{"type":70,"value":544},{"type":64,"tag":361,"props":1157,"children":1158},{"style":401},[1159],{"type":70,"value":1116},{"type":64,"tag":361,"props":1161,"children":1163},{"class":363,"line":1162},8,[1164,1169],{"type":64,"tag":361,"props":1165,"children":1166},{"style":390},[1167],{"type":70,"value":1168},"  --local-dir",{"type":64,"tag":361,"props":1170,"children":1171},{"style":390},[1172],{"type":70,"value":1173}," .\u002Fgr00t_dataset\n",{"type":64,"tag":361,"props":1175,"children":1177},{"class":363,"line":1176},9,[1178],{"type":64,"tag":361,"props":1179,"children":1180},{"emptyLinePlaceholder":890},[1181],{"type":70,"value":893},{"type":64,"tag":361,"props":1183,"children":1185},{"class":363,"line":1184},10,[1186],{"type":64,"tag":361,"props":1187,"children":1188},{"style":368},[1189],{"type":70,"value":1190},"# Sparse-checkout via git-LFS (if you want incremental git-style work)\n",{"type":64,"tag":361,"props":1192,"children":1194},{"class":363,"line":1193},11,[1195,1199,1204,1209,1214,1219,1223,1227,1231],{"type":64,"tag":361,"props":1196,"children":1197},{"style":499},[1198],{"type":70,"value":280},{"type":64,"tag":361,"props":1200,"children":1201},{"style":390},[1202],{"type":70,"value":1203}," clone",{"type":64,"tag":361,"props":1205,"children":1206},{"style":390},[1207],{"type":70,"value":1208}," --filter=blob:none",{"type":64,"tag":361,"props":1210,"children":1211},{"style":390},[1212],{"type":70,"value":1213}," --no-checkout",{"type":64,"tag":361,"props":1215,"children":1216},{"style":390},[1217],{"type":70,"value":1218}," https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002F",{"type":64,"tag":361,"props":1220,"children":1221},{"style":384},[1222],{"type":70,"value":1028},{"type":64,"tag":361,"props":1224,"children":1225},{"style":390},[1226],{"type":70,"value":1033},{"type":64,"tag":361,"props":1228,"children":1229},{"style":401},[1230],{"type":70,"value":1038},{"type":64,"tag":361,"props":1232,"children":1233},{"style":384},[1234],{"type":70,"value":1080},{"type":64,"tag":361,"props":1236,"children":1238},{"class":363,"line":1237},12,[1239,1244,1249,1253,1257],{"type":64,"tag":361,"props":1240,"children":1241},{"style":378},[1242],{"type":70,"value":1243},"cd",{"type":64,"tag":361,"props":1245,"children":1246},{"style":384},[1247],{"type":70,"value":1248}," \u003C",{"type":64,"tag":361,"props":1250,"children":1251},{"style":390},[1252],{"type":70,"value":1033},{"type":64,"tag":361,"props":1254,"children":1255},{"style":401},[1256],{"type":70,"value":1038},{"type":64,"tag":361,"props":1258,"children":1259},{"style":384},[1260],{"type":70,"value":1080},{"type":64,"tag":361,"props":1262,"children":1264},{"class":363,"line":1263},13,[1265,1269,1274,1279],{"type":64,"tag":361,"props":1266,"children":1267},{"style":499},[1268],{"type":70,"value":280},{"type":64,"tag":361,"props":1270,"children":1271},{"style":390},[1272],{"type":70,"value":1273}," sparse-checkout",{"type":64,"tag":361,"props":1275,"children":1276},{"style":390},[1277],{"type":70,"value":1278}," init",{"type":64,"tag":361,"props":1280,"children":1281},{"style":390},[1282],{"type":70,"value":1283}," --cone\n",{"type":64,"tag":361,"props":1285,"children":1287},{"class":363,"line":1286},14,[1288,1292,1296,1301,1305,1310,1315],{"type":64,"tag":361,"props":1289,"children":1290},{"style":499},[1291],{"type":70,"value":280},{"type":64,"tag":361,"props":1293,"children":1294},{"style":390},[1295],{"type":70,"value":1273},{"type":64,"tag":361,"props":1297,"children":1298},{"style":390},[1299],{"type":70,"value":1300}," set",{"type":64,"tag":361,"props":1302,"children":1303},{"style":384},[1304],{"type":70,"value":1248},{"type":64,"tag":361,"props":1306,"children":1307},{"style":390},[1308],{"type":70,"value":1309},"subfolde",{"type":64,"tag":361,"props":1311,"children":1312},{"style":401},[1313],{"type":70,"value":1314},"r",{"type":64,"tag":361,"props":1316,"children":1317},{"style":384},[1318],{"type":70,"value":1080},{"type":64,"tag":361,"props":1320,"children":1322},{"class":363,"line":1321},15,[1323,1327,1332],{"type":64,"tag":361,"props":1324,"children":1325},{"style":499},[1326],{"type":70,"value":280},{"type":64,"tag":361,"props":1328,"children":1329},{"style":390},[1330],{"type":70,"value":1331}," checkout",{"type":64,"tag":361,"props":1333,"children":1334},{"style":390},[1335],{"type":70,"value":1336}," main\n",{"type":64,"tag":80,"props":1338,"children":1339},{},[1340],{"type":70,"value":1341},"Special-case downloaders:",{"type":64,"tag":152,"props":1343,"children":1344},{},[1345,1505,1539,1557],{"type":64,"tag":156,"props":1346,"children":1347},{},[1348,1352,1354,1360,1362,1368,1370,1374,1387,1389,1395,1397,1402,1404,1410,1412,1417,1419,1422,1424,1429,1431,1437,1439,1442,1444,1448,1450],{"type":64,"tag":129,"props":1349,"children":1350},{},[1351],{"type":70,"value":667},{"type":70,"value":1353}," (133 TB) — use the official Python\ntoolkit ",{"type":64,"tag":86,"props":1355,"children":1357},{"className":1356},[],[1358],{"type":70,"value":1359},"pip install physical_ai_av",{"type":70,"value":1361},"\n(",{"type":64,"tag":206,"props":1363,"children":1365},{"href":59,"rel":1364},[209],[1366],{"type":70,"value":1367},"NVlabs\u002Fphysical_ai_av",{"type":70,"value":1369},")\nto filter by sensor \u002F country \u002F split before downloading; otherwise\nyou will pull TBs you don't need.",{"type":64,"tag":1371,"props":1372,"children":1373},"br",{},[],{"type":64,"tag":129,"props":1375,"children":1376},{},[1377,1379,1385],{"type":70,"value":1378},"Default policy: pre-filter to ",{"type":64,"tag":86,"props":1380,"children":1382},{"className":1381},[],[1383],{"type":70,"value":1384},"platform_class == 'hyperion_8.1'",{"type":70,"value":1386},"\nbefore any bulk pull.",{"type":70,"value":1388}," The dataset mixes two sensor rigs\n(",{"type":64,"tag":86,"props":1390,"children":1392},{"className":1391},[],[1393],{"type":70,"value":1394},"hyperion_8",{"type":70,"value":1396}," and ",{"type":64,"tag":86,"props":1398,"children":1400},{"className":1399},[],[1401],{"type":70,"value":659},{"type":70,"value":1403},") and the entire downstream\nNuRec \u002F NCore \u002F Asset-Harvester chain in this repo is ",{"type":64,"tag":1405,"props":1406,"children":1407},"em",{},[1408],{"type":70,"value":1409},"only",{"type":70,"value":1411},"\nvalidated for ",{"type":64,"tag":86,"props":1413,"children":1415},{"className":1414},[],[1416],{"type":70,"value":659},{"type":70,"value":1418},". Roughly half of the 306k clips fall\noutside that platform, so the filter typically cuts the working set\n(and download size) in half.",{"type":64,"tag":1371,"props":1420,"children":1421},{},[],{"type":70,"value":1423},"Apply this filter by default in any download script you generate.\nOnly skip the filter when the user ",{"type":64,"tag":129,"props":1425,"children":1426},{},[1427],{"type":70,"value":1428},"explicitly",{"type":70,"value":1430}," asks for it (e.g.\n\"download all platforms\", \"include hyperion_8\", \"don't filter by\nplatform\") or when they hand you a specific clip UUID — in that\nsingle-clip case, ",{"type":64,"tag":86,"props":1432,"children":1434},{"className":1433},[],[1435],{"type":70,"value":1436},"download_clip_features(clip_id=...)",{"type":70,"value":1438}," is already\nscoped and no platform filter is needed.",{"type":64,"tag":1371,"props":1440,"children":1441},{},[],{"type":70,"value":1443},"Canonical filtered recipe (see § ",{"type":64,"tag":206,"props":1445,"children":1446},{"href":648},[1447],{"type":70,"value":651},{"type":70,"value":1449},"\nbelow for an end-to-end example):",{"type":64,"tag":351,"props":1451,"children":1454},{"className":1452,"code":1453,"language":46,"meta":355,"style":355},"language-python shiki shiki-themes material-theme-lighter material-theme material-theme-palenight","from physical_ai_av import PhysicalAIAVDatasetInterface\n\ndataset = PhysicalAIAVDatasetInterface()\ndataset.download_metadata()\ndc = dataset.metadata['data_collection']\nhyperion_81_clip_ids = dc[dc['platform_class'] == 'hyperion_8.1'].index.tolist()\n",[1455],{"type":64,"tag":86,"props":1456,"children":1457},{"__ignoreMap":355},[1458,1466,1473,1481,1489,1497],{"type":64,"tag":361,"props":1459,"children":1460},{"class":363,"line":364},[1461],{"type":64,"tag":361,"props":1462,"children":1463},{},[1464],{"type":70,"value":1465},"from physical_ai_av import PhysicalAIAVDatasetInterface\n",{"type":64,"tag":361,"props":1467,"children":1468},{"class":363,"line":374},[1469],{"type":64,"tag":361,"props":1470,"children":1471},{"emptyLinePlaceholder":890},[1472],{"type":70,"value":893},{"type":64,"tag":361,"props":1474,"children":1475},{"class":363,"line":886},[1476],{"type":64,"tag":361,"props":1477,"children":1478},{},[1479],{"type":70,"value":1480},"dataset = PhysicalAIAVDatasetInterface()\n",{"type":64,"tag":361,"props":1482,"children":1483},{"class":363,"line":30},[1484],{"type":64,"tag":361,"props":1485,"children":1486},{},[1487],{"type":70,"value":1488},"dataset.download_metadata()\n",{"type":64,"tag":361,"props":1490,"children":1491},{"class":363,"line":935},[1492],{"type":64,"tag":361,"props":1493,"children":1494},{},[1495],{"type":70,"value":1496},"dc = dataset.metadata['data_collection']\n",{"type":64,"tag":361,"props":1498,"children":1499},{"class":363,"line":1119},[1500],{"type":64,"tag":361,"props":1501,"children":1502},{},[1503],{"type":70,"value":1504},"hyperion_81_clip_ids = dc[dc['platform_class'] == 'hyperion_8.1'].index.tolist()\n",{"type":64,"tag":156,"props":1506,"children":1507},{},[1508,1513,1515,1521,1522,1529,1531,1537],{"type":64,"tag":129,"props":1509,"children":1510},{},[1511],{"type":70,"value":1512},"PhysicalAI-Autonomous-Vehicle-Cosmos-Drive-Dreams",{"type":70,"value":1514}," (3 TB) — use the\nupstream ",{"type":64,"tag":86,"props":1516,"children":1518},{"className":1517},[],[1519],{"type":70,"value":1520},"download.py",{"type":70,"value":1361},{"type":64,"tag":206,"props":1523,"children":1526},{"href":1524,"rel":1525},"https:\u002F\u002Fgithub.com\u002Fnv-tlabs\u002FCosmos-Drive-Dreams\u002Fblob\u002Fmain\u002Fscripts\u002Fdownload.py",[209],[1527],{"type":70,"value":1528},"nv-tlabs\u002FCosmos-Drive-Dreams",{"type":70,"value":1530},")\nwith ",{"type":64,"tag":86,"props":1532,"children":1534},{"className":1533},[],[1535],{"type":70,"value":1536},"--file_types {hdmap,lidar,synthetic}",{"type":70,"value":1538}," to select layers.",{"type":64,"tag":156,"props":1540,"children":1541},{},[1542,1547,1549,1555],{"type":64,"tag":129,"props":1543,"children":1544},{},[1545],{"type":70,"value":1546},"PhysicalAI-SpatialIntelligence-Lyra-SDG",{"type":70,"value":1548}," (25 TB) — ",{"type":64,"tag":86,"props":1550,"children":1552},{"className":1551},[],[1553],{"type":70,"value":1554},"hf download --local-dir lyra_dataset\u002Ftar",{"type":70,"value":1556},"; untar each tar yourself.",{"type":64,"tag":156,"props":1558,"children":1559},{},[1560,1565],{"type":64,"tag":129,"props":1561,"children":1562},{},[1563],{"type":70,"value":1564},"Spatial-Intelligence-Warehouse",{"type":70,"value":1566}," — chunked TAR-GZs need a manual\nloop after download (script provided in the upstream card).",{"type":64,"tag":80,"props":1568,"children":1569},{},[1570,1572,1578],{"type":70,"value":1571},"For dataset filtering \u002F preview: NVIDIA's\n",{"type":64,"tag":206,"props":1573,"children":1575},{"href":58,"rel":1574},[209],[1576],{"type":70,"value":1577},"Cosmos Dataset Search (CDS)",{"type":70,"value":1579},"\nlets you query a 41K subset of the AV dataset semantically before\ndownloading.",{"type":64,"tag":73,"props":1581,"children":1583},{"id":1582},"filtered-av-download-recipe",[1584],{"type":70,"value":651},{"type":64,"tag":80,"props":1586,"children":1587},{},[1588,1590,1595,1597,1602,1604,1609,1611,1616,1617,1622,1623,1628,1629,1634],{"type":70,"value":1589},"Always use this recipe (or a derivative of it) when the user asks to\ndownload raw clips from ",{"type":64,"tag":86,"props":1591,"children":1593},{"className":1592},[],[1594],{"type":70,"value":667},{"type":70,"value":1596},". The\n",{"type":64,"tag":86,"props":1598,"children":1600},{"className":1599},[],[1601],{"type":70,"value":659},{"type":70,"value":1603}," filter is the ",{"type":64,"tag":129,"props":1605,"children":1606},{},[1607],{"type":70,"value":1608},"default",{"type":70,"value":1610},"; it matches every downstream\nskill in this repo (",{"type":64,"tag":86,"props":1612,"children":1614},{"className":1613},[],[1615],{"type":70,"value":99},{"type":70,"value":101},{"type":64,"tag":86,"props":1618,"children":1620},{"className":1619},[],[1621],{"type":70,"value":107},{"type":70,"value":101},{"type":64,"tag":86,"props":1624,"children":1626},{"className":1625},[],[1627],{"type":70,"value":115},{"type":70,"value":101},{"type":64,"tag":86,"props":1630,"children":1632},{"className":1631},[],[1633],{"type":70,"value":122},{"type":70,"value":271},{"type":64,"tag":80,"props":1636,"children":1637},{},[1638,1640,1645],{"type":70,"value":1639},"When to ",{"type":64,"tag":129,"props":1641,"children":1642},{},[1643],{"type":70,"value":1644},"skip",{"type":70,"value":1646}," the platform filter:",{"type":64,"tag":629,"props":1648,"children":1649},{},[1650,1662],{"type":64,"tag":156,"props":1651,"children":1652},{},[1653,1655,1660],{"type":70,"value":1654},"The user explicitly opts out — e.g. \"download all platforms\",\n\"include hyperion_8 too\", \"ignore platform_class\", or asks for a\ndataset-wide statistic. In that case, drop the platform mask and\nwarn them that NuRec \u002F NCore tooling will not work on the\n",{"type":64,"tag":86,"props":1656,"children":1658},{"className":1657},[],[1659],{"type":70,"value":1394},{"type":70,"value":1661}," clips.",{"type":64,"tag":156,"props":1663,"children":1664},{},[1665,1667,1672,1674,1679,1681,1687],{"type":70,"value":1666},"The user gave you a ",{"type":64,"tag":129,"props":1668,"children":1669},{},[1670],{"type":70,"value":1671},"specific clip UUID",{"type":70,"value":1673},". A single-clip\n",{"type":64,"tag":86,"props":1675,"children":1677},{"className":1676},[],[1678],{"type":70,"value":1436},{"type":70,"value":1680}," is already scoped, and\nfiltering by ",{"type":64,"tag":86,"props":1682,"children":1684},{"className":1683},[],[1685],{"type":70,"value":1686},"platform_class",{"type":70,"value":1688}," for one clip is pointless.",{"type":64,"tag":80,"props":1690,"children":1691},{},[1692],{"type":70,"value":1693},"Recipe (defaults: filter on; user may layer extra masks like country\nor sensor presence on top):",{"type":64,"tag":351,"props":1695,"children":1697},{"className":1452,"code":1696,"language":46,"meta":355,"style":355},"from physical_ai_av import PhysicalAIAVDatasetInterface\n\ndataset = PhysicalAIAVDatasetInterface()\n\ndataset.download_metadata()\ndc = dataset.metadata['data_collection']\n\nclip_mask = dc['platform_class'] == 'hyperion_8.1'\n\n# Optional extra masks (only add when the user asked for them):\n# clip_mask &= dc['country'] == 'US'\n# sp = dataset.metadata['feature_presence']   # 26.03+; was 'sensor_presence' in 25.10\n# clip_mask &= sp['lidar_top_360fov']\n\nclip_ids = dc[clip_mask].index.tolist()\nprint(f\"Downloading {len(clip_ids)} hyperion_8.1 clips\")\n\ndataset.download_clip_features(\n    clip_id=clip_ids,\n    features=[\"camera_front_wide_120fov\", \"lidar_top_360fov\", \"egomotion\"],\n    max_workers=8,\n)\n",[1698],{"type":64,"tag":86,"props":1699,"children":1700},{"__ignoreMap":355},[1701,1708,1715,1722,1729,1736,1743,1750,1758,1765,1773,1781,1789,1797,1804,1812,1820,1828,1837,1846,1855,1864],{"type":64,"tag":361,"props":1702,"children":1703},{"class":363,"line":364},[1704],{"type":64,"tag":361,"props":1705,"children":1706},{},[1707],{"type":70,"value":1465},{"type":64,"tag":361,"props":1709,"children":1710},{"class":363,"line":374},[1711],{"type":64,"tag":361,"props":1712,"children":1713},{"emptyLinePlaceholder":890},[1714],{"type":70,"value":893},{"type":64,"tag":361,"props":1716,"children":1717},{"class":363,"line":886},[1718],{"type":64,"tag":361,"props":1719,"children":1720},{},[1721],{"type":70,"value":1480},{"type":64,"tag":361,"props":1723,"children":1724},{"class":363,"line":30},[1725],{"type":64,"tag":361,"props":1726,"children":1727},{"emptyLinePlaceholder":890},[1728],{"type":70,"value":893},{"type":64,"tag":361,"props":1730,"children":1731},{"class":363,"line":935},[1732],{"type":64,"tag":361,"props":1733,"children":1734},{},[1735],{"type":70,"value":1488},{"type":64,"tag":361,"props":1737,"children":1738},{"class":363,"line":1119},[1739],{"type":64,"tag":361,"props":1740,"children":1741},{},[1742],{"type":70,"value":1496},{"type":64,"tag":361,"props":1744,"children":1745},{"class":363,"line":1136},[1746],{"type":64,"tag":361,"props":1747,"children":1748},{"emptyLinePlaceholder":890},[1749],{"type":70,"value":893},{"type":64,"tag":361,"props":1751,"children":1752},{"class":363,"line":1162},[1753],{"type":64,"tag":361,"props":1754,"children":1755},{},[1756],{"type":70,"value":1757},"clip_mask = dc['platform_class'] == 'hyperion_8.1'\n",{"type":64,"tag":361,"props":1759,"children":1760},{"class":363,"line":1176},[1761],{"type":64,"tag":361,"props":1762,"children":1763},{"emptyLinePlaceholder":890},[1764],{"type":70,"value":893},{"type":64,"tag":361,"props":1766,"children":1767},{"class":363,"line":1184},[1768],{"type":64,"tag":361,"props":1769,"children":1770},{},[1771],{"type":70,"value":1772},"# Optional extra masks (only add when the user asked for them):\n",{"type":64,"tag":361,"props":1774,"children":1775},{"class":363,"line":1193},[1776],{"type":64,"tag":361,"props":1777,"children":1778},{},[1779],{"type":70,"value":1780},"# clip_mask &= dc['country'] == 'US'\n",{"type":64,"tag":361,"props":1782,"children":1783},{"class":363,"line":1237},[1784],{"type":64,"tag":361,"props":1785,"children":1786},{},[1787],{"type":70,"value":1788},"# sp = dataset.metadata['feature_presence']   # 26.03+; was 'sensor_presence' in 25.10\n",{"type":64,"tag":361,"props":1790,"children":1791},{"class":363,"line":1263},[1792],{"type":64,"tag":361,"props":1793,"children":1794},{},[1795],{"type":70,"value":1796},"# clip_mask &= sp['lidar_top_360fov']\n",{"type":64,"tag":361,"props":1798,"children":1799},{"class":363,"line":1286},[1800],{"type":64,"tag":361,"props":1801,"children":1802},{"emptyLinePlaceholder":890},[1803],{"type":70,"value":893},{"type":64,"tag":361,"props":1805,"children":1806},{"class":363,"line":1321},[1807],{"type":64,"tag":361,"props":1808,"children":1809},{},[1810],{"type":70,"value":1811},"clip_ids = dc[clip_mask].index.tolist()\n",{"type":64,"tag":361,"props":1813,"children":1814},{"class":363,"line":26},[1815],{"type":64,"tag":361,"props":1816,"children":1817},{},[1818],{"type":70,"value":1819},"print(f\"Downloading {len(clip_ids)} hyperion_8.1 clips\")\n",{"type":64,"tag":361,"props":1821,"children":1823},{"class":363,"line":1822},17,[1824],{"type":64,"tag":361,"props":1825,"children":1826},{"emptyLinePlaceholder":890},[1827],{"type":70,"value":893},{"type":64,"tag":361,"props":1829,"children":1831},{"class":363,"line":1830},18,[1832],{"type":64,"tag":361,"props":1833,"children":1834},{},[1835],{"type":70,"value":1836},"dataset.download_clip_features(\n",{"type":64,"tag":361,"props":1838,"children":1840},{"class":363,"line":1839},19,[1841],{"type":64,"tag":361,"props":1842,"children":1843},{},[1844],{"type":70,"value":1845},"    clip_id=clip_ids,\n",{"type":64,"tag":361,"props":1847,"children":1849},{"class":363,"line":1848},20,[1850],{"type":64,"tag":361,"props":1851,"children":1852},{},[1853],{"type":70,"value":1854},"    features=[\"camera_front_wide_120fov\", \"lidar_top_360fov\", \"egomotion\"],\n",{"type":64,"tag":361,"props":1856,"children":1858},{"class":363,"line":1857},21,[1859],{"type":64,"tag":361,"props":1860,"children":1861},{},[1862],{"type":70,"value":1863},"    max_workers=8,\n",{"type":64,"tag":361,"props":1865,"children":1867},{"class":363,"line":1866},22,[1868],{"type":64,"tag":361,"props":1869,"children":1870},{},[1871],{"type":70,"value":1872},")\n",{"type":64,"tag":80,"props":1874,"children":1875},{},[1876],{"type":70,"value":1877},"Single-clip fast path (no platform filter — the UUID is already\nspecific):",{"type":64,"tag":351,"props":1879,"children":1881},{"className":1452,"code":1880,"language":46,"meta":355,"style":355},"dataset.download_clip_features(\n    clip_id=\"\u003Cpaste-clip-uuid>\",\n    features=[\"camera_front_wide_120fov\", \"lidar_top_360fov\", \"egomotion\"],\n)\n",[1882],{"type":64,"tag":86,"props":1883,"children":1884},{"__ignoreMap":355},[1885,1892,1900,1907],{"type":64,"tag":361,"props":1886,"children":1887},{"class":363,"line":364},[1888],{"type":64,"tag":361,"props":1889,"children":1890},{},[1891],{"type":70,"value":1836},{"type":64,"tag":361,"props":1893,"children":1894},{"class":363,"line":374},[1895],{"type":64,"tag":361,"props":1896,"children":1897},{},[1898],{"type":70,"value":1899},"    clip_id=\"\u003Cpaste-clip-uuid>\",\n",{"type":64,"tag":361,"props":1901,"children":1902},{"class":363,"line":886},[1903],{"type":64,"tag":361,"props":1904,"children":1905},{},[1906],{"type":70,"value":1854},{"type":64,"tag":361,"props":1908,"children":1909},{"class":363,"line":30},[1910],{"type":64,"tag":361,"props":1911,"children":1912},{},[1913],{"type":70,"value":1872},{"type":64,"tag":80,"props":1915,"children":1916},{},[1917],{"type":70,"value":1918},"Explicit opt-out (only when the user asked for it):",{"type":64,"tag":351,"props":1920,"children":1922},{"className":1452,"code":1921,"language":46,"meta":355,"style":355},"clip_ids = dc.index.tolist()   # NO platform filter — all 306k clips\n# WARNING: downstream NuRec \u002F NCore \u002F Asset-Harvester only handle\n# the hyperion_8.1 subset.\n",[1923],{"type":64,"tag":86,"props":1924,"children":1925},{"__ignoreMap":355},[1926,1934,1942],{"type":64,"tag":361,"props":1927,"children":1928},{"class":363,"line":364},[1929],{"type":64,"tag":361,"props":1930,"children":1931},{},[1932],{"type":70,"value":1933},"clip_ids = dc.index.tolist()   # NO platform filter — all 306k clips\n",{"type":64,"tag":361,"props":1935,"children":1936},{"class":363,"line":374},[1937],{"type":64,"tag":361,"props":1938,"children":1939},{},[1940],{"type":70,"value":1941},"# WARNING: downstream NuRec \u002F NCore \u002F Asset-Harvester only handle\n",{"type":64,"tag":361,"props":1943,"children":1944},{"class":363,"line":886},[1945],{"type":64,"tag":361,"props":1946,"children":1947},{},[1948],{"type":70,"value":1949},"# the hyperion_8.1 subset.\n",{"type":64,"tag":73,"props":1951,"children":1953},{"id":1952},"pick-a-dataset-by-task",[1954],{"type":70,"value":678},{"type":64,"tag":1956,"props":1957,"children":1958},"table",{},[1959,1978],{"type":64,"tag":1960,"props":1961,"children":1962},"thead",{},[1963],{"type":64,"tag":1964,"props":1965,"children":1966},"tr",{},[1967,1973],{"type":64,"tag":1968,"props":1969,"children":1970},"th",{},[1971],{"type":70,"value":1972},"Goal",{"type":64,"tag":1968,"props":1974,"children":1975},{},[1976],{"type":70,"value":1977},"Recommended dataset(s)",{"type":64,"tag":1979,"props":1980,"children":1981},"tbody",{},[1982,2001,2031,2049,2068,2084,2103,2122,2146,2165,2184,2203,2220,2244,2261,2278,2297,2314,2333,2350,2369,2388,2407,2425,2444,2463,2482],{"type":64,"tag":1964,"props":1983,"children":1984},{},[1985,1991],{"type":64,"tag":1986,"props":1987,"children":1988},"td",{},[1989],{"type":70,"value":1990},"End-to-end AV training (real, multi-sensor)",{"type":64,"tag":1986,"props":1992,"children":1993},{},[1994,1999],{"type":64,"tag":86,"props":1995,"children":1997},{"className":1996},[],[1998],{"type":70,"value":667},{"type":70,"value":2000}," (133 TB, 1700 h, 25 countries)",{"type":64,"tag":1964,"props":2002,"children":2003},{},[2004,2020],{"type":64,"tag":1986,"props":2005,"children":2006},{},[2007,2009,2018],{"type":70,"value":2008},"AV in NCore V4 format (drop-in for ",{"type":64,"tag":206,"props":2010,"children":2012},{"href":2011},"..\u002Fncore\u002FSKILL.md",[2013],{"type":64,"tag":86,"props":2014,"children":2016},{"className":2015},[],[2017],{"type":70,"value":99},{"type":70,"value":2019},")",{"type":64,"tag":1986,"props":2021,"children":2022},{},[2023,2029],{"type":64,"tag":86,"props":2024,"children":2026},{"className":2025},[],[2027],{"type":70,"value":2028},"PhysicalAI-Autonomous-Vehicles-NCore",{"type":70,"value":2030}," (~1.1k clips)",{"type":64,"tag":1964,"props":2032,"children":2033},{},[2034,2039],{"type":64,"tag":1986,"props":2035,"children":2036},{},[2037],{"type":70,"value":2038},"AV photoreal Sim2Real \u002F weather augmentation",{"type":64,"tag":1986,"props":2040,"children":2041},{},[2042,2047],{"type":64,"tag":86,"props":2043,"children":2045},{"className":2044},[],[2046],{"type":70,"value":1512},{"type":70,"value":2048}," (3 TB; 7 weather variants)",{"type":64,"tag":1964,"props":2050,"children":2051},{},[2052,2057],{"type":64,"tag":1986,"props":2053,"children":2054},{},[2055],{"type":70,"value":2056},"AV neural reconstructions ready for CARLA \u002F NuRec",{"type":64,"tag":1986,"props":2058,"children":2059},{},[2060,2066],{"type":64,"tag":86,"props":2061,"children":2063},{"className":2062},[],[2064],{"type":70,"value":2065},"PhysicalAI-Autonomous-Vehicles-NuRec",{"type":70,"value":2067}," (918 USDZ scenes)",{"type":64,"tag":1964,"props":2069,"children":2070},{},[2071,2076],{"type":64,"tag":1986,"props":2072,"children":2073},{},[2074],{"type":70,"value":2075},"Asset Harvester \u002F 3DGS extraction sample clip",{"type":64,"tag":1986,"props":2077,"children":2078},{},[2079],{"type":64,"tag":86,"props":2080,"children":2082},{"className":2081},[],[2083],{"type":70,"value":2028},{"type":64,"tag":1964,"props":2085,"children":2086},{},[2087,2092],{"type":64,"tag":1986,"props":2088,"children":2089},{},[2090],{"type":70,"value":2091},"GR00T post-training, broad coverage",{"type":64,"tag":1986,"props":2093,"children":2094},{},[2095,2101],{"type":64,"tag":86,"props":2096,"children":2098},{"className":2097},[],[2099],{"type":70,"value":2100},"PhysicalAI-Robotics-GR00T-X-Embodiment-Sim",{"type":70,"value":2102}," (1.91 TB, 24 GR1 task families + bimanual + RoboCasa)",{"type":64,"tag":1964,"props":2104,"children":2105},{},[2106,2111],{"type":64,"tag":1986,"props":2107,"children":2108},{},[2109],{"type":70,"value":2110},"GR00T fine-tune on industrial tasks",{"type":64,"tag":1986,"props":2112,"children":2113},{},[2114,2120],{"type":64,"tag":86,"props":2115,"children":2117},{"className":2116},[],[2118],{"type":70,"value":2119},"PhysicalAI-GR00T-Tuned-Tasks",{"type":70,"value":2121}," (Nut Pouring, Exhaust Pipe Sorting)",{"type":64,"tag":1964,"props":2123,"children":2124},{},[2125,2130],{"type":64,"tag":1986,"props":2126,"children":2127},{},[2128],{"type":70,"value":2129},"GR00T eval images \u002F videos",{"type":64,"tag":1986,"props":2131,"children":2132},{},[2133,2139,2140],{"type":64,"tag":86,"props":2134,"children":2136},{"className":2135},[],[2137],{"type":70,"value":2138},"PhysicalAI-Robotics-GR00T-Eval",{"type":70,"value":101},{"type":64,"tag":86,"props":2141,"children":2143},{"className":2142},[],[2144],{"type":70,"value":2145},"PhysicalAI-Robotics-GR00T-GR1",{"type":64,"tag":1964,"props":2147,"children":2148},{},[2149,2154],{"type":64,"tag":1986,"props":2150,"children":2151},{},[2152],{"type":70,"value":2153},"Real humanoid teleop (Unitree G1)",{"type":64,"tag":1986,"props":2155,"children":2156},{},[2157,2163],{"type":64,"tag":86,"props":2158,"children":2160},{"className":2159},[],[2161],{"type":70,"value":2162},"PhysicalAI-Robotics-GR00T-Teleop-G1",{"type":70,"value":2164}," (1000 trajectories)",{"type":64,"tag":1964,"props":2166,"children":2167},{},[2168,2173],{"type":64,"tag":1986,"props":2169,"children":2170},{},[2171],{"type":70,"value":2172},"Sim humanoid teleop (Fourier GR1)",{"type":64,"tag":1986,"props":2174,"children":2175},{},[2176,2182],{"type":64,"tag":86,"props":2177,"children":2179},{"className":2178},[],[2180],{"type":70,"value":2181},"PhysicalAI-Robotics-GR00T-Teleop-Sim",{"type":70,"value":2183}," (24 tasks × 1k trajectories)",{"type":64,"tag":1964,"props":2185,"children":2186},{},[2187,2192],{"type":64,"tag":1986,"props":2188,"children":2189},{},[2190],{"type":70,"value":2191},"Massive humanoid pretraining (44k h, DreamDojo)",{"type":64,"tag":1986,"props":2193,"children":2194},{},[2195,2201],{"type":64,"tag":86,"props":2196,"children":2198},{"className":2197},[],[2199],{"type":70,"value":2200},"PhysicalAI-Robotics-GR00T-Teleop-GR1",{"type":70,"value":2202}," (74.3 GB)",{"type":64,"tag":1964,"props":2204,"children":2205},{},[2206,2211],{"type":64,"tag":1986,"props":2207,"children":2208},{},[2209],{"type":70,"value":2210},"Spatial-memory imitation learning (mindmap)",{"type":64,"tag":1986,"props":2212,"children":2213},{},[2214],{"type":64,"tag":86,"props":2215,"children":2217},{"className":2216},[],[2218],{"type":70,"value":2219},"PhysicalAI-Robotics-mindmap-{Stick-in-Bin,Drill-in-Box,Cube-Stacking,Mug-in-Drawer}",{"type":64,"tag":1964,"props":2221,"children":2222},{},[2223,2228],{"type":64,"tag":1986,"props":2224,"children":2225},{},[2226],{"type":70,"value":2227},"Robot pick-place in kitchen (bimanual Kinova Gen3)",{"type":64,"tag":1986,"props":2229,"children":2230},{},[2231,2237,2238],{"type":64,"tag":86,"props":2232,"children":2234},{"className":2233},[],[2235],{"type":70,"value":2236},"PhysicalAI-Robotics-Manipulation-Kitchen",{"type":70,"value":101},{"type":64,"tag":86,"props":2239,"children":2241},{"className":2240},[],[2242],{"type":70,"value":2243},"-Manipulation-Objects",{"type":64,"tag":1964,"props":2245,"children":2246},{},[2247,2252],{"type":64,"tag":1986,"props":2248,"children":2249},{},[2250],{"type":70,"value":2251},"Robot pick-place tabletop (single Franka)",{"type":64,"tag":1986,"props":2253,"children":2254},{},[2255],{"type":64,"tag":86,"props":2256,"children":2258},{"className":2257},[],[2259],{"type":70,"value":2260},"PhysicalAI-Robotics-Manipulation-SingleArm",{"type":64,"tag":1964,"props":2262,"children":2263},{},[2264,2269],{"type":64,"tag":1986,"props":2265,"children":2266},{},[2267],{"type":70,"value":2268},"Cosmos-Transfer1 visual-augmented stacking",{"type":64,"tag":1986,"props":2270,"children":2271},{},[2272],{"type":64,"tag":86,"props":2273,"children":2275},{"className":2274},[],[2276],{"type":70,"value":2277},"PhysicalAI-Robotics-Manipulation-Augmented",{"type":64,"tag":1964,"props":2279,"children":2280},{},[2281,2286],{"type":64,"tag":1986,"props":2282,"children":2283},{},[2284],{"type":70,"value":2285},"Massive teleop in kitchen (Franka + mobile base)",{"type":64,"tag":1986,"props":2287,"children":2288},{},[2289,2295],{"type":64,"tag":86,"props":2290,"children":2292},{"className":2291},[],[2293],{"type":70,"value":2294},"PhysicalAI-Robotics-Manipulation-Kitchen-Demos",{"type":70,"value":2296}," (600 h, 316 tasks, 55k traj)",{"type":64,"tag":1964,"props":2298,"children":2299},{},[2300,2305],{"type":64,"tag":1986,"props":2301,"children":2302},{},[2303],{"type":70,"value":2304},"MJCF kitchen objects + fixtures (MuJoCo)",{"type":64,"tag":1986,"props":2306,"children":2307},{},[2308],{"type":64,"tag":86,"props":2309,"children":2311},{"className":2310},[],[2312],{"type":70,"value":2313},"PhysicalAI-Robotics-Manipulation-Objects-Kitchen-MJCF",{"type":64,"tag":1964,"props":2315,"children":2316},{},[2317,2322],{"type":64,"tag":1986,"props":2318,"children":2319},{},[2320],{"type":70,"value":2321},"Sim-Ready warehouse for IsaacSim",{"type":64,"tag":1986,"props":2323,"children":2324},{},[2325,2331],{"type":64,"tag":86,"props":2326,"children":2328},{"className":2327},[],[2329],{"type":70,"value":2330},"PhysicalAI-SimReady-Warehouse-01",{"type":70,"value":2332}," (753 USD assets)",{"type":64,"tag":1964,"props":2334,"children":2335},{},[2336,2341],{"type":64,"tag":1986,"props":2337,"children":2338},{},[2339],{"type":70,"value":2340},"GR1 tabletop digital cousins (assets)",{"type":64,"tag":1986,"props":2342,"children":2343},{},[2344],{"type":64,"tag":86,"props":2345,"children":2347},{"className":2346},[],[2348],{"type":70,"value":2349},"PhysicalAI-DigitalCousin-Assets",{"type":64,"tag":1964,"props":2351,"children":2352},{},[2353,2358],{"type":64,"tag":1986,"props":2354,"children":2355},{},[2356],{"type":70,"value":2357},"3DGS \u002F Sim-Ready indoor scenes for AMR sim",{"type":64,"tag":1986,"props":2359,"children":2360},{},[2361,2367],{"type":64,"tag":86,"props":2362,"children":2364},{"className":2363},[],[2365],{"type":70,"value":2366},"PhysicalAI-Robotics-NuRec",{"type":70,"value":2368}," (Nova Carter labs, Zurich offices, hand-held)",{"type":64,"tag":1964,"props":2370,"children":2371},{},[2372,2377],{"type":64,"tag":1986,"props":2373,"children":2374},{},[2375],{"type":70,"value":2376},"Multi-cam tracking + 3D box benchmark",{"type":64,"tag":1986,"props":2378,"children":2379},{},[2380,2386],{"type":64,"tag":86,"props":2381,"children":2383},{"className":2382},[],[2384],{"type":70,"value":2385},"PhysicalAI-SmartSpaces",{"type":70,"value":2387}," (AI City Challenge 2024 + 2025)",{"type":64,"tag":1964,"props":2389,"children":2390},{},[2391,2396],{"type":64,"tag":1986,"props":2392,"children":2393},{},[2394],{"type":70,"value":2395},"3D scene QA \u002F VLM training (warehouses)",{"type":64,"tag":1986,"props":2397,"children":2398},{},[2399,2405],{"type":64,"tag":86,"props":2400,"children":2402},{"className":2401},[],[2403],{"type":70,"value":2404},"PhysicalAI-Spatial-Intelligence-Warehouse",{"type":70,"value":2406}," (499k QA pairs)",{"type":64,"tag":1964,"props":2408,"children":2409},{},[2410,2415],{"type":64,"tag":1986,"props":2411,"children":2412},{},[2413],{"type":70,"value":2414},"Generative 3D scene reconstruction training",{"type":64,"tag":1986,"props":2416,"children":2417},{},[2418,2423],{"type":64,"tag":86,"props":2419,"children":2421},{"className":2420},[],[2422],{"type":70,"value":1546},{"type":70,"value":2424}," (25 TB; GEN3C-derived)",{"type":64,"tag":1964,"props":2426,"children":2427},{},[2428,2433],{"type":64,"tag":1986,"props":2429,"children":2430},{},[2431],{"type":70,"value":2432},"Radiance-field photometric benchmark",{"type":64,"tag":1986,"props":2434,"children":2435},{},[2436,2442],{"type":64,"tag":86,"props":2437,"children":2439},{"className":2438},[],[2440],{"type":70,"value":2441},"PhysicalAI-NuRec-PPISP",{"type":70,"value":2443}," (8 sequences, +\u002F-2 EV bracketing)",{"type":64,"tag":1964,"props":2445,"children":2446},{},[2447,2452],{"type":64,"tag":1986,"props":2448,"children":2449},{},[2450],{"type":70,"value":2451},"Grasping models (Franka, Robotiq-2f-140, suction)",{"type":64,"tag":1986,"props":2453,"children":2454},{},[2455,2461],{"type":64,"tag":86,"props":2456,"children":2458},{"className":2457},[],[2459],{"type":70,"value":2460},"PhysicalAI-Robotics-GraspGen",{"type":70,"value":2462}," (57M grasps, Objaverse-LVIS)",{"type":64,"tag":1964,"props":2464,"children":2465},{},[2466,2471],{"type":64,"tag":1986,"props":2467,"children":2468},{},[2469],{"type":70,"value":2470},"Healthcare \u002F surgical robotics autonomy",{"type":64,"tag":1986,"props":2472,"children":2473},{},[2474,2480],{"type":64,"tag":86,"props":2475,"children":2477},{"className":2476},[],[2478],{"type":70,"value":2479},"PhysicalAI-Robotics-Open-H-Embodiment",{"type":70,"value":2481}," (750 h, 4.5 TB, 30+ orgs)",{"type":64,"tag":1964,"props":2483,"children":2484},{},[2485,2490],{"type":64,"tag":1986,"props":2486,"children":2487},{},[2488],{"type":70,"value":2489},"Volumetric mechanical \u002F material properties",{"type":64,"tag":1986,"props":2491,"children":2492},{},[2493,2499,2500],{"type":64,"tag":86,"props":2494,"children":2496},{"className":2495},[],[2497],{"type":70,"value":2498},"PhysicalAI-Robotics-PhysicalAssets-VoMP",{"type":70,"value":101},{"type":64,"tag":86,"props":2501,"children":2503},{"className":2502},[],[2504],{"type":70,"value":2505},"-VoMP-Eval",{"type":64,"tag":73,"props":2507,"children":2508},{"id":55},[2509],{"type":70,"value":689},{"type":64,"tag":319,"props":2511,"children":2513},{"id":2512},"physicalai-autonomous-vehicles",[2514],{"type":70,"value":667},{"type":64,"tag":80,"props":2516,"children":2517},{},[2518],{"type":70,"value":2519},"The flagship real-world AV dataset.",{"type":64,"tag":1956,"props":2521,"children":2522},{},[2523,2539],{"type":64,"tag":1960,"props":2524,"children":2525},{},[2526],{"type":64,"tag":1964,"props":2527,"children":2528},{},[2529,2534],{"type":64,"tag":1968,"props":2530,"children":2531},{},[2532],{"type":70,"value":2533},"Field",{"type":64,"tag":1968,"props":2535,"children":2536},{},[2537],{"type":70,"value":2538},"Value",{"type":64,"tag":1979,"props":2540,"children":2541},{},[2542,2559,2582,2595,2608,2621,2639,2678,2691,2710],{"type":64,"tag":1964,"props":2543,"children":2544},{},[2545,2550],{"type":64,"tag":1986,"props":2546,"children":2547},{},[2548],{"type":70,"value":2549},"HF",{"type":64,"tag":1986,"props":2551,"children":2552},{},[2553],{"type":64,"tag":206,"props":2554,"children":2557},{"href":2555,"rel":2556},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Autonomous-Vehicles",[209],[2558],{"type":70,"value":2555},{"type":64,"tag":1964,"props":2560,"children":2561},{},[2562,2567],{"type":64,"tag":1986,"props":2563,"children":2564},{},[2565],{"type":70,"value":2566},"Size",{"type":64,"tag":1986,"props":2568,"children":2569},{},[2570,2575,2577],{"type":64,"tag":129,"props":2571,"children":2572},{},[2573],{"type":70,"value":2574},"133 TB",{"type":70,"value":2576}," — 306,152 clips × 20 s = ",{"type":64,"tag":129,"props":2578,"children":2579},{},[2580],{"type":70,"value":2581},"1700 h",{"type":64,"tag":1964,"props":2583,"children":2584},{},[2585,2590],{"type":64,"tag":1986,"props":2586,"children":2587},{},[2588],{"type":70,"value":2589},"Sensors",{"type":64,"tag":1986,"props":2591,"children":2592},{},[2593],{"type":70,"value":2594},"7 cameras (1080p @ 30 FPS), top-360 LiDAR (10 Hz, 298k clips), up to 10 radars (160k clips), ego motion, calibration, machine labels",{"type":64,"tag":1964,"props":2596,"children":2597},{},[2598,2603],{"type":64,"tag":1986,"props":2599,"children":2600},{},[2601],{"type":70,"value":2602},"Geography",{"type":64,"tag":1986,"props":2604,"children":2605},{},[2606],{"type":70,"value":2607},"25 countries, 2500+ cities (US 155k, Germany 44k, …)",{"type":64,"tag":1964,"props":2609,"children":2610},{},[2611,2616],{"type":64,"tag":1986,"props":2612,"children":2613},{},[2614],{"type":70,"value":2615},"Format",{"type":64,"tag":1986,"props":2617,"children":2618},{},[2619],{"type":70,"value":2620},"Per-sensor parquet\u002Fmp4 chunks of ~100 clips; UUIDs cross-link sensors",{"type":64,"tag":1964,"props":2622,"children":2623},{},[2624,2629],{"type":64,"tag":1986,"props":2625,"children":2626},{},[2627],{"type":70,"value":2628},"License",{"type":64,"tag":1986,"props":2630,"children":2631},{},[2632,2637],{"type":64,"tag":129,"props":2633,"children":2634},{},[2635],{"type":70,"value":2636},"NVIDIA AV Dataset License Agreement",{"type":70,"value":2638}," (gated; AV-development-only purpose; no biometric \u002F surveillance \u002F re-identification; expires 12 months after download)",{"type":64,"tag":1964,"props":2640,"children":2641},{},[2642,2647],{"type":64,"tag":1986,"props":2643,"children":2644},{},[2645],{"type":70,"value":2646},"Toolkit",{"type":64,"tag":1986,"props":2648,"children":2649},{},[2650,2655,2657,2670,2672,2676],{"type":64,"tag":86,"props":2651,"children":2653},{"className":2652},[],[2654],{"type":70,"value":1359},{"type":70,"value":2656}," — direct filtered downloads + format docs. ",{"type":64,"tag":129,"props":2658,"children":2659},{},[2660,2662,2668],{"type":70,"value":2661},"Default to filtering ",{"type":64,"tag":86,"props":2663,"children":2665},{"className":2664},[],[2666],{"type":70,"value":2667},"metadata['data_collection']['platform_class'] == 'hyperion_8.1'",{"type":70,"value":2669}," before any bulk pull",{"type":70,"value":2671}," (see ",{"type":64,"tag":206,"props":2673,"children":2674},{"href":648},[2675],{"type":70,"value":651},{"type":70,"value":2677},"); only skip the filter on explicit user request or when downloading a specific clip UUID.",{"type":64,"tag":1964,"props":2679,"children":2680},{},[2681,2686],{"type":64,"tag":1986,"props":2682,"children":2683},{},[2684],{"type":70,"value":2685},"Versions",{"type":64,"tag":1986,"props":2687,"children":2688},{},[2689],{"type":70,"value":2690},"26.03 (current; offline-optimized features for 97 % of clips), 25.10 (initial)",{"type":64,"tag":1964,"props":2692,"children":2693},{},[2694,2699],{"type":64,"tag":1986,"props":2695,"children":2696},{},[2697],{"type":70,"value":2698},"Subset preview",{"type":64,"tag":1986,"props":2700,"children":2701},{},[2702,2704],{"type":70,"value":2703},"41k clips searchable on ",{"type":64,"tag":206,"props":2705,"children":2707},{"href":58,"rel":2706},[209],[2708],{"type":70,"value":2709},"Cosmos Dataset Search",{"type":64,"tag":1964,"props":2711,"children":2712},{},[2713,2718],{"type":64,"tag":1986,"props":2714,"children":2715},{},[2716],{"type":70,"value":2717},"Use with",{"type":64,"tag":1986,"props":2719,"children":2720},{},[2721,2729,2731,2740,2742,2748,2750,2756,2757,2763],{"type":64,"tag":206,"props":2722,"children":2723},{"href":2011},[2724],{"type":64,"tag":86,"props":2725,"children":2727},{"className":2726},[],[2728],{"type":70,"value":2011},{"type":70,"value":2730}," (convert raw clips to NCore V4), ",{"type":64,"tag":206,"props":2732,"children":2734},{"href":2733},"..\u002Fasset-harvester\u002FSKILL.md",[2735],{"type":64,"tag":86,"props":2736,"children":2738},{"className":2737},[],[2739],{"type":70,"value":2733},{"type":70,"value":2741}," (extract per-object Gaussian assets). NuRec workflows are only validated for ",{"type":64,"tag":86,"props":2743,"children":2745},{"className":2744},[],[2746],{"type":70,"value":2747},"platform_class == hyperion_8.1",{"type":70,"value":2749},". Upstream sim\u002Ftraining tools without an in-repo skill: ",{"type":64,"tag":86,"props":2751,"children":2753},{"className":2752},[],[2754],{"type":70,"value":2755},"NVlabs\u002Falpamayo-1.5",{"type":70,"value":101},{"type":64,"tag":86,"props":2758,"children":2760},{"className":2759},[],[2761],{"type":70,"value":2762},"NVlabs\u002Falpasim",{"type":70,"value":2764},", CARLA.",{"type":64,"tag":319,"props":2766,"children":2768},{"id":2767},"physicalai-autonomous-vehicles-ncore",[2769],{"type":70,"value":2028},{"type":64,"tag":80,"props":2771,"children":2772},{},[2773],{"type":70,"value":2774},"Curated NCore V4 subset of the above.",{"type":64,"tag":1956,"props":2776,"children":2777},{},[2778,2792],{"type":64,"tag":1960,"props":2779,"children":2780},{},[2781],{"type":64,"tag":1964,"props":2782,"children":2783},{},[2784,2788],{"type":64,"tag":1968,"props":2785,"children":2786},{},[2787],{"type":70,"value":2533},{"type":64,"tag":1968,"props":2789,"children":2790},{},[2791],{"type":70,"value":2538},{"type":64,"tag":1979,"props":2793,"children":2794},{},[2795,2811,2823,2851,2863,2900],{"type":64,"tag":1964,"props":2796,"children":2797},{},[2798,2802],{"type":64,"tag":1986,"props":2799,"children":2800},{},[2801],{"type":70,"value":2549},{"type":64,"tag":1986,"props":2803,"children":2804},{},[2805],{"type":64,"tag":206,"props":2806,"children":2809},{"href":2807,"rel":2808},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Autonomous-Vehicles-NCore",[209],[2810],{"type":70,"value":2807},{"type":64,"tag":1964,"props":2812,"children":2813},{},[2814,2818],{"type":64,"tag":1986,"props":2815,"children":2816},{},[2817],{"type":70,"value":2566},{"type":64,"tag":1986,"props":2819,"children":2820},{},[2821],{"type":70,"value":2822},"~1.1k clips with accurate offline calibration \u002F egomotion \u002F cuboids",{"type":64,"tag":1964,"props":2824,"children":2825},{},[2826,2830],{"type":64,"tag":1986,"props":2827,"children":2828},{},[2829],{"type":70,"value":2615},{"type":64,"tag":1986,"props":2831,"children":2832},{},[2833,2835,2841,2843,2849],{"type":70,"value":2834},"NCore V4 — ",{"type":64,"tag":86,"props":2836,"children":2838},{"className":2837},[],[2839],{"type":70,"value":2840},"pai_\u003Cuuid>.json",{"type":70,"value":2842}," + per-sensor ",{"type":64,"tag":86,"props":2844,"children":2846},{"className":2845},[],[2847],{"type":70,"value":2848},".zarr.itar",{"type":70,"value":2850}," files",{"type":64,"tag":1964,"props":2852,"children":2853},{},[2854,2858],{"type":64,"tag":1986,"props":2855,"children":2856},{},[2857],{"type":70,"value":2628},{"type":64,"tag":1986,"props":2859,"children":2860},{},[2861],{"type":70,"value":2862},"NVIDIA AV Dataset License Agreement (gated, same as above)",{"type":64,"tag":1964,"props":2864,"children":2865},{},[2866,2870],{"type":64,"tag":1986,"props":2867,"children":2868},{},[2869],{"type":70,"value":2717},{"type":64,"tag":1986,"props":2871,"children":2872},{},[2873,2881,2883,2891,2893,2899],{"type":64,"tag":206,"props":2874,"children":2875},{"href":2011},[2876],{"type":64,"tag":86,"props":2877,"children":2879},{"className":2878},[],[2880],{"type":70,"value":2011},{"type":70,"value":2882}," (drop-in), ",{"type":64,"tag":206,"props":2884,"children":2885},{"href":2733},[2886],{"type":64,"tag":86,"props":2887,"children":2889},{"className":2888},[],[2890],{"type":70,"value":2733},{"type":70,"value":2892}," (sample clip path: ",{"type":64,"tag":86,"props":2894,"children":2896},{"className":2895},[],[2897],{"type":70,"value":2898},"clips\u002F2a6f330-5ab0-4e92-99d4-d19e406952f4\u002F",{"type":70,"value":2019},{"type":64,"tag":1964,"props":2901,"children":2902},{},[2903,2908],{"type":64,"tag":1986,"props":2904,"children":2905},{},[2906],{"type":70,"value":2907},"Notes",{"type":64,"tag":1986,"props":2909,"children":2910},{},[2911,2913,2920],{"type":70,"value":2912},"Built via ",{"type":64,"tag":206,"props":2914,"children":2917},{"href":2915,"rel":2916},"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fncore\u002Ftree\u002Fmain\u002Ftools\u002Fdata_converter\u002Fpai",[209],[2918],{"type":70,"value":2919},"PAI data converter",{"type":70,"value":2921},". Use this BEFORE the full AV dataset for any NCore-driven workflow.",{"type":64,"tag":319,"props":2923,"children":2925},{"id":2924},"physicalai-autonomous-vehicles-nurec",[2926],{"type":70,"value":2065},{"type":64,"tag":80,"props":2928,"children":2929},{},[2930],{"type":70,"value":2931},"Pre-built NuRec dynamic neural reconstructions ready for IsaacSim \u002F CARLA.",{"type":64,"tag":1956,"props":2933,"children":2934},{},[2935,2949],{"type":64,"tag":1960,"props":2936,"children":2937},{},[2938],{"type":64,"tag":1964,"props":2939,"children":2940},{},[2941,2945],{"type":64,"tag":1968,"props":2942,"children":2943},{},[2944],{"type":70,"value":2533},{"type":64,"tag":1968,"props":2946,"children":2947},{},[2948],{"type":70,"value":2538},{"type":64,"tag":1979,"props":2950,"children":2951},{},[2952,2968,2988,3001,3013,3025],{"type":64,"tag":1964,"props":2953,"children":2954},{},[2955,2959],{"type":64,"tag":1986,"props":2956,"children":2957},{},[2958],{"type":70,"value":2549},{"type":64,"tag":1986,"props":2960,"children":2961},{},[2962],{"type":64,"tag":206,"props":2963,"children":2966},{"href":2964,"rel":2965},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Autonomous-Vehicles-NuRec",[209],[2967],{"type":70,"value":2964},{"type":64,"tag":1964,"props":2969,"children":2970},{},[2971,2975],{"type":64,"tag":1986,"props":2972,"children":2973},{},[2974],{"type":70,"value":2566},{"type":64,"tag":1986,"props":2976,"children":2977},{},[2978,2980,2986],{"type":70,"value":2979},"918 USDZ scenes, ~20 s each, with surface meshes + front-camera mp4 + ",{"type":64,"tag":86,"props":2981,"children":2983},{"className":2982},[],[2984],{"type":70,"value":2985},"labels.json",{"type":70,"value":2987}," (Batch0002+)",{"type":64,"tag":1964,"props":2989,"children":2990},{},[2991,2996],{"type":64,"tag":1986,"props":2992,"children":2993},{},[2994],{"type":70,"value":2995},"Reconstruction",{"type":64,"tag":1986,"props":2997,"children":2998},{},[2999],{"type":70,"value":3000},"6 cameras (front-wide 120°, front-tele 30°, cross-L\u002FR 120°, rear-L\u002FR 70°)",{"type":64,"tag":1964,"props":3002,"children":3003},{},[3004,3008],{"type":64,"tag":1986,"props":3005,"children":3006},{},[3007],{"type":70,"value":2685},{"type":64,"tag":1986,"props":3009,"children":3010},{},[3011],{"type":70,"value":3012},"26.02 (current), 25.07, 25.05",{"type":64,"tag":1964,"props":3014,"children":3015},{},[3016,3020],{"type":64,"tag":1986,"props":3017,"children":3018},{},[3019],{"type":70,"value":2628},{"type":64,"tag":1986,"props":3021,"children":3022},{},[3023],{"type":70,"value":3024},"NVIDIA AV Dataset License Agreement (gated)",{"type":64,"tag":1964,"props":3026,"children":3027},{},[3028,3032],{"type":64,"tag":1986,"props":3029,"children":3030},{},[3031],{"type":70,"value":2717},{"type":64,"tag":1986,"props":3033,"children":3034},{},[3035,3044,3046,3052,3054,3063],{"type":64,"tag":206,"props":3036,"children":3038},{"href":3037},"..\u002Fnre\u002FSKILL.md",[3039],{"type":64,"tag":86,"props":3040,"children":3042},{"className":3041},[],[3043],{"type":70,"value":3037},{"type":70,"value":3045}," (render the USDZs locally or over ",{"type":64,"tag":86,"props":3047,"children":3049},{"className":3048},[],[3050],{"type":70,"value":3051},"serve-grpc",{"type":70,"value":3053},"), ",{"type":64,"tag":206,"props":3055,"children":3057},{"href":3056},"..\u002Fnurec-fixer\u002FSKILL.md",[3058],{"type":64,"tag":86,"props":3059,"children":3061},{"className":3060},[],[3062],{"type":70,"value":3056},{"type":70,"value":3064}," (clean up rendered frames). Upstream consumer without an in-repo skill: CARLA (NuRec integration in 0.9.16+).",{"type":64,"tag":319,"props":3066,"children":3068},{"id":3067},"physicalai-autonomous-vehicle-cosmos-drive-dreams",[3069],{"type":70,"value":1512},{"type":64,"tag":80,"props":3071,"children":3072},{},[3073],{"type":70,"value":3074},"Cosmos-Transfer-style synthetic + HD-map labels for diverse weather.",{"type":64,"tag":1956,"props":3076,"children":3077},{},[3078,3092],{"type":64,"tag":1960,"props":3079,"children":3080},{},[3081],{"type":64,"tag":1964,"props":3082,"children":3083},{},[3084,3088],{"type":64,"tag":1968,"props":3085,"children":3086},{},[3087],{"type":70,"value":2533},{"type":64,"tag":1968,"props":3089,"children":3090},{},[3091],{"type":70,"value":2538},{"type":64,"tag":1979,"props":3093,"children":3094},{},[3095,3111,3123,3136,3149,3162,3179,3196,3213],{"type":64,"tag":1964,"props":3096,"children":3097},{},[3098,3102],{"type":64,"tag":1986,"props":3099,"children":3100},{},[3101],{"type":70,"value":2549},{"type":64,"tag":1986,"props":3103,"children":3104},{},[3105],{"type":64,"tag":206,"props":3106,"children":3109},{"href":3107,"rel":3108},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Autonomous-Vehicle-Cosmos-Drive-Dreams",[209],[3110],{"type":70,"value":3107},{"type":64,"tag":1964,"props":3112,"children":3113},{},[3114,3118],{"type":64,"tag":1986,"props":3115,"children":3116},{},[3117],{"type":70,"value":2566},{"type":64,"tag":1986,"props":3119,"children":3120},{},[3121],{"type":70,"value":3122},"3 TB total (synthetic only ~700 GB) — 5,843 RDS-HQ clips × 2 chunks × 7 weather = 81,802 synthetic videos (121 frames each)",{"type":64,"tag":1964,"props":3124,"children":3125},{},[3126,3131],{"type":64,"tag":1986,"props":3127,"children":3128},{},[3129],{"type":70,"value":3130},"Modalities",{"type":64,"tag":1986,"props":3132,"children":3133},{},[3134],{"type":70,"value":3135},"Cosmos-generated MP4, HDMap (lanes\u002Flanelines\u002Froad boundaries\u002Fwait lines\u002Fcrosswalks\u002Fmarkings\u002Fpoles\u002Flights\u002Fsigns), LiDAR, vehicle pose, camera intrinsics (ftheta + pinhole), 4D object tracking",{"type":64,"tag":1964,"props":3137,"children":3138},{},[3139,3144],{"type":64,"tag":1986,"props":3140,"children":3141},{},[3142],{"type":70,"value":3143},"Cameras",{"type":64,"tag":1986,"props":3145,"children":3146},{},[3147],{"type":70,"value":3148},"7 (front-wide\u002Fcross-L\u002Fcross-R\u002Frear-L\u002Frear-R\u002Frear-tele\u002Ffront-tele)",{"type":64,"tag":1964,"props":3150,"children":3151},{},[3152,3157],{"type":64,"tag":1986,"props":3153,"children":3154},{},[3155],{"type":70,"value":3156},"Weather variants",{"type":64,"tag":1986,"props":3158,"children":3159},{},[3160],{"type":70,"value":3161},"Foggy \u002F Golden hour \u002F Morning \u002F Night \u002F Rainy \u002F Snowy \u002F Sunny",{"type":64,"tag":1964,"props":3163,"children":3164},{},[3165,3169],{"type":64,"tag":1986,"props":3166,"children":3167},{},[3168],{"type":70,"value":2628},{"type":64,"tag":1986,"props":3170,"children":3171},{},[3172,3177],{"type":64,"tag":129,"props":3173,"children":3174},{},[3175],{"type":70,"value":3176},"CC-BY-4.0",{"type":70,"value":3178}," (commercial OK with attribution)",{"type":64,"tag":1964,"props":3180,"children":3181},{},[3182,3187],{"type":64,"tag":1986,"props":3183,"children":3184},{},[3185],{"type":70,"value":3186},"Tooling",{"type":64,"tag":1986,"props":3188,"children":3189},{},[3190],{"type":64,"tag":86,"props":3191,"children":3193},{"className":3192},[],[3194],{"type":70,"value":3195},"wget … scripts\u002Fdownload.py; python download.py --odir \u003Cpath> --file_types hdmap,lidar,synthetic",{"type":64,"tag":1964,"props":3197,"children":3198},{},[3199,3204],{"type":64,"tag":1986,"props":3200,"children":3201},{},[3202],{"type":70,"value":3203},"Paper",{"type":64,"tag":1986,"props":3205,"children":3206},{},[3207],{"type":64,"tag":206,"props":3208,"children":3211},{"href":3209,"rel":3210},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2506.09042",[209],[3212],{"type":70,"value":3209},{"type":64,"tag":1964,"props":3214,"children":3215},{},[3216,3220],{"type":64,"tag":1986,"props":3217,"children":3218},{},[3219],{"type":70,"value":2717},{"type":64,"tag":1986,"props":3221,"children":3222},{},[3223,3225,3231,3232,3238,3239,3244],{"type":70,"value":3224},"Upstream consumers without an in-repo skill: ",{"type":64,"tag":86,"props":3226,"children":3228},{"className":3227},[],[3229],{"type":70,"value":3230},"nvidia\u002FCosmos-Transfer1",{"type":70,"value":101},{"type":64,"tag":86,"props":3233,"children":3235},{"className":3234},[],[3236],{"type":70,"value":3237},"nvidia\u002FCosmos-Predict",{"type":70,"value":101},{"type":64,"tag":86,"props":3240,"children":3242},{"className":3241},[],[3243],{"type":70,"value":2762},{"type":70,"value":2764},{"type":64,"tag":319,"props":3246,"children":3248},{"id":3247},"physicalai-autonomous-vehicle-cosmos-synthetic",[3249],{"type":70,"value":3250},"PhysicalAI-Autonomous-Vehicle-Cosmos-Synthetic",{"type":64,"tag":1956,"props":3252,"children":3253},{},[3254,3268],{"type":64,"tag":1960,"props":3255,"children":3256},{},[3257],{"type":64,"tag":1964,"props":3258,"children":3259},{},[3260,3264],{"type":64,"tag":1968,"props":3261,"children":3262},{},[3263],{"type":70,"value":2533},{"type":64,"tag":1968,"props":3265,"children":3266},{},[3267],{"type":70,"value":2538},{"type":64,"tag":1979,"props":3269,"children":3270},{},[3271,3287],{"type":64,"tag":1964,"props":3272,"children":3273},{},[3274,3278],{"type":64,"tag":1986,"props":3275,"children":3276},{},[3277],{"type":70,"value":2549},{"type":64,"tag":1986,"props":3279,"children":3280},{},[3281],{"type":64,"tag":206,"props":3282,"children":3285},{"href":3283,"rel":3284},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Autonomous-Vehicle-Cosmos-Synthetic",[209],[3286],{"type":70,"value":3283},{"type":64,"tag":1964,"props":3288,"children":3289},{},[3290,3295],{"type":64,"tag":1986,"props":3291,"children":3292},{},[3293],{"type":70,"value":3294},"Status",{"type":64,"tag":1986,"props":3296,"children":3297},{},[3298,3303,3305,3310],{"type":64,"tag":129,"props":3299,"children":3300},{},[3301],{"type":70,"value":3302},"Pointer \u002F placeholder",{"type":70,"value":3304}," — content moved to ",{"type":64,"tag":86,"props":3306,"children":3308},{"className":3307},[],[3309],{"type":70,"value":1512},{"type":70,"value":3311},". Use that. (Card is 2.59 kB.)",{"type":64,"tag":73,"props":3313,"children":3315},{"id":3314},"robotics-manipulation",[3316],{"type":70,"value":700},{"type":64,"tag":80,"props":3318,"children":3319},{},[3320,3322,3327,3329,3335],{"type":70,"value":3321},"All in ",{"type":64,"tag":129,"props":3323,"children":3324},{},[3325],{"type":70,"value":3326},"LeRobot v2.x",{"type":70,"value":3328}," format unless noted, generated in IsaacSim with\ntask-and-motion planning + ",{"type":64,"tag":86,"props":3330,"children":3332},{"className":3331},[],[3333],{"type":70,"value":3334},"scene_synthesizer",{"type":70,"value":3336}," procedural scenes +\nCuRobo motion generation.",{"type":64,"tag":319,"props":3338,"children":3340},{"id":3339},"physicalai-robotics-manipulation-kitchen",[3341],{"type":70,"value":2236},{"type":64,"tag":80,"props":3343,"children":3344},{},[3345],{"type":70,"value":3346},"Bimanual Kinova Gen3 in procedurally-generated kitchens.",{"type":64,"tag":1956,"props":3348,"children":3349},{},[3350,3364],{"type":64,"tag":1960,"props":3351,"children":3352},{},[3353],{"type":64,"tag":1964,"props":3354,"children":3355},{},[3356,3360],{"type":64,"tag":1968,"props":3357,"children":3358},{},[3359],{"type":70,"value":2533},{"type":64,"tag":1968,"props":3361,"children":3362},{},[3363],{"type":70,"value":2538},{"type":64,"tag":1979,"props":3365,"children":3366},{},[3367,3383,3395,3408,3421,3433,3444],{"type":64,"tag":1964,"props":3368,"children":3369},{},[3370,3374],{"type":64,"tag":1986,"props":3371,"children":3372},{},[3373],{"type":70,"value":2549},{"type":64,"tag":1986,"props":3375,"children":3376},{},[3377],{"type":64,"tag":206,"props":3378,"children":3381},{"href":3379,"rel":3380},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-Manipulation-Kitchen",[209],[3382],{"type":70,"value":3379},{"type":64,"tag":1964,"props":3384,"children":3385},{},[3386,3390],{"type":64,"tag":1986,"props":3387,"children":3388},{},[3389],{"type":70,"value":2566},{"type":64,"tag":1986,"props":3391,"children":3392},{},[3393],{"type":70,"value":3394},"12 GB total",{"type":64,"tag":1964,"props":3396,"children":3397},{},[3398,3403],{"type":64,"tag":1986,"props":3399,"children":3400},{},[3401],{"type":70,"value":3402},"Tasks",{"type":64,"tag":1986,"props":3404,"children":3405},{},[3406],{"type":70,"value":3407},"open\u002Fclose × {cabinet, dishwasher, fridge, drawer} = 8",{"type":64,"tag":1964,"props":3409,"children":3410},{},[3411,3416],{"type":64,"tag":1986,"props":3412,"children":3413},{},[3414],{"type":70,"value":3415},"Trajectories",{"type":64,"tag":1986,"props":3417,"children":3418},{},[3419],{"type":70,"value":3420},"~874 episodes total (range 72–205 per task)",{"type":64,"tag":1964,"props":3422,"children":3423},{},[3424,3428],{"type":64,"tag":1986,"props":3425,"children":3426},{},[3427],{"type":70,"value":3143},{"type":64,"tag":1986,"props":3429,"children":3430},{},[3431],{"type":70,"value":3432},"6 × 512² RGB+depth+segmentation (world \u002F external \u002F each wrist \u002F head)",{"type":64,"tag":1964,"props":3434,"children":3435},{},[3436,3440],{"type":64,"tag":1986,"props":3437,"children":3438},{},[3439],{"type":70,"value":2628},{"type":64,"tag":1986,"props":3441,"children":3442},{},[3443],{"type":70,"value":3176},{"type":64,"tag":1964,"props":3445,"children":3446},{},[3447,3452],{"type":64,"tag":1986,"props":3448,"children":3449},{},[3450],{"type":70,"value":3451},"Commercial",{"type":64,"tag":1986,"props":3453,"children":3454},{},[3455],{"type":70,"value":3456},"✅",{"type":64,"tag":319,"props":3458,"children":3460},{"id":3459},"physicalai-robotics-manipulation-objects",[3461],{"type":70,"value":3462},"PhysicalAI-Robotics-Manipulation-Objects",{"type":64,"tag":80,"props":3464,"children":3465},{},[3466],{"type":70,"value":3467},"Same kitchen environment, bimanual Kinova; pick \u002F place bench \u002F place cabinet.",{"type":64,"tag":1956,"props":3469,"children":3470},{},[3471,3485],{"type":64,"tag":1960,"props":3472,"children":3473},{},[3474],{"type":64,"tag":1964,"props":3475,"children":3476},{},[3477,3481],{"type":64,"tag":1968,"props":3478,"children":3479},{},[3480],{"type":70,"value":2533},{"type":64,"tag":1968,"props":3482,"children":3483},{},[3484],{"type":70,"value":2538},{"type":64,"tag":1979,"props":3486,"children":3487},{},[3488,3504,3516,3548,3560],{"type":64,"tag":1964,"props":3489,"children":3490},{},[3491,3495],{"type":64,"tag":1986,"props":3492,"children":3493},{},[3494],{"type":70,"value":2549},{"type":64,"tag":1986,"props":3496,"children":3497},{},[3498],{"type":64,"tag":206,"props":3499,"children":3502},{"href":3500,"rel":3501},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-Manipulation-Objects",[209],[3503],{"type":70,"value":3500},{"type":64,"tag":1964,"props":3505,"children":3506},{},[3507,3511],{"type":64,"tag":1986,"props":3508,"children":3509},{},[3510],{"type":70,"value":2566},{"type":64,"tag":1986,"props":3512,"children":3513},{},[3514],{"type":70,"value":3515},"4.26 GB",{"type":64,"tag":1964,"props":3517,"children":3518},{},[3519,3523],{"type":64,"tag":1986,"props":3520,"children":3521},{},[3522],{"type":70,"value":3402},{"type":64,"tag":1986,"props":3524,"children":3525},{},[3526,3532,3533,3539,3540,3546],{"type":64,"tag":86,"props":3527,"children":3529},{"className":3528},[],[3530],{"type":70,"value":3531},"pick",{"type":70,"value":101},{"type":64,"tag":86,"props":3534,"children":3536},{"className":3535},[],[3537],{"type":70,"value":3538},"place_bench",{"type":70,"value":101},{"type":64,"tag":86,"props":3541,"children":3543},{"className":3542},[],[3544],{"type":70,"value":3545},"place_cabinet",{"type":70,"value":3547}," (540 episodes total)",{"type":64,"tag":1964,"props":3549,"children":3550},{},[3551,3555],{"type":64,"tag":1986,"props":3552,"children":3553},{},[3554],{"type":70,"value":2628},{"type":64,"tag":1986,"props":3556,"children":3557},{},[3558],{"type":70,"value":3559},"CC-BY-4.0 (intended R&D only per card)",{"type":64,"tag":1964,"props":3561,"children":3562},{},[3563,3567],{"type":64,"tag":1986,"props":3564,"children":3565},{},[3566],{"type":70,"value":2717},{"type":64,"tag":1986,"props":3568,"children":3569},{},[3570],{"type":70,"value":3571},"Upstream Isaac Sim \u002F Isaac Lab (no in-repo skill).",{"type":64,"tag":319,"props":3573,"children":3575},{"id":3574},"physicalai-robotics-manipulation-singlearm",[3576],{"type":70,"value":2260},{"type":64,"tag":80,"props":3578,"children":3579},{},[3580],{"type":70,"value":3581},"Franka Panda tabletop, procedurally generated.",{"type":64,"tag":1956,"props":3583,"children":3584},{},[3585,3599],{"type":64,"tag":1960,"props":3586,"children":3587},{},[3588],{"type":64,"tag":1964,"props":3589,"children":3590},{},[3591,3595],{"type":64,"tag":1968,"props":3592,"children":3593},{},[3594],{"type":70,"value":2533},{"type":64,"tag":1968,"props":3596,"children":3597},{},[3598],{"type":70,"value":2538},{"type":64,"tag":1979,"props":3600,"children":3601},{},[3602,3618,3630,3676,3688,3701],{"type":64,"tag":1964,"props":3603,"children":3604},{},[3605,3609],{"type":64,"tag":1986,"props":3606,"children":3607},{},[3608],{"type":70,"value":2549},{"type":64,"tag":1986,"props":3610,"children":3611},{},[3612],{"type":64,"tag":206,"props":3613,"children":3616},{"href":3614,"rel":3615},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-Manipulation-SingleArm",[209],[3617],{"type":70,"value":3614},{"type":64,"tag":1964,"props":3619,"children":3620},{},[3621,3625],{"type":64,"tag":1986,"props":3622,"children":3623},{},[3624],{"type":70,"value":2566},{"type":64,"tag":1986,"props":3626,"children":3627},{},[3628],{"type":70,"value":3629},"15.3 GB",{"type":64,"tag":1964,"props":3631,"children":3632},{},[3633,3637],{"type":64,"tag":1986,"props":3634,"children":3635},{},[3636],{"type":70,"value":3402},{"type":64,"tag":1986,"props":3638,"children":3639},{},[3640,3646,3647,3653,3654,3660,3661,3667,3668,3674],{"type":64,"tag":86,"props":3641,"children":3643},{"className":3642},[],[3644],{"type":70,"value":3645},"panda-stack-wide",{"type":70,"value":101},{"type":64,"tag":86,"props":3648,"children":3650},{"className":3649},[],[3651],{"type":70,"value":3652},"panda-stack-platforms",{"type":70,"value":101},{"type":64,"tag":86,"props":3655,"children":3657},{"className":3656},[],[3658],{"type":70,"value":3659},"panda-stack-platforms-texture",{"type":70,"value":101},{"type":64,"tag":86,"props":3662,"children":3664},{"className":3663},[],[3665],{"type":70,"value":3666},"panda-open-cabinet-{left,right}",{"type":70,"value":101},{"type":64,"tag":86,"props":3669,"children":3671},{"className":3670},[],[3672],{"type":70,"value":3673},"panda-open-drawer",{"type":70,"value":3675}," (~38k episodes)",{"type":64,"tag":1964,"props":3677,"children":3678},{},[3679,3683],{"type":64,"tag":1986,"props":3680,"children":3681},{},[3682],{"type":70,"value":3130},{"type":64,"tag":1986,"props":3684,"children":3685},{},[3686],{"type":70,"value":3687},"World cam + wrist cam (RGB + depth on the texture\u002Fcabinet\u002Fdrawer subsets)",{"type":64,"tag":1964,"props":3689,"children":3690},{},[3691,3696],{"type":64,"tag":1986,"props":3692,"children":3693},{},[3694],{"type":70,"value":3695},"State",{"type":64,"tag":1986,"props":3697,"children":3698},{},[3699],{"type":70,"value":3700},"53 D (stack-wide) \u002F 81 D (others) — proprioception + object poses",{"type":64,"tag":1964,"props":3702,"children":3703},{},[3704,3708],{"type":64,"tag":1986,"props":3705,"children":3706},{},[3707],{"type":70,"value":2628},{"type":64,"tag":1986,"props":3709,"children":3710},{},[3711],{"type":70,"value":3712},"CC-BY-4.0; commercial OK",{"type":64,"tag":319,"props":3714,"children":3716},{"id":3715},"physicalai-robotics-manipulation-augmented",[3717],{"type":70,"value":2277},{"type":64,"tag":80,"props":3719,"children":3720},{},[3721],{"type":70,"value":3722},"Mimic-generated Franka cube-stacking, plus Cosmos-Transfer1 visual augmentation.",{"type":64,"tag":1956,"props":3724,"children":3725},{},[3726,3740],{"type":64,"tag":1960,"props":3727,"children":3728},{},[3729],{"type":64,"tag":1964,"props":3730,"children":3731},{},[3732,3736],{"type":64,"tag":1968,"props":3733,"children":3734},{},[3735],{"type":70,"value":2533},{"type":64,"tag":1968,"props":3737,"children":3738},{},[3739],{"type":70,"value":2538},{"type":64,"tag":1979,"props":3741,"children":3742},{},[3743,3759,3771,3784,3797,3808,3826],{"type":64,"tag":1964,"props":3744,"children":3745},{},[3746,3750],{"type":64,"tag":1986,"props":3747,"children":3748},{},[3749],{"type":70,"value":2549},{"type":64,"tag":1986,"props":3751,"children":3752},{},[3753],{"type":64,"tag":206,"props":3754,"children":3757},{"href":3755,"rel":3756},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-Manipulation-Augmented",[209],[3758],{"type":70,"value":3755},{"type":64,"tag":1964,"props":3760,"children":3761},{},[3762,3766],{"type":64,"tag":1986,"props":3763,"children":3764},{},[3765],{"type":70,"value":2566},{"type":64,"tag":1986,"props":3767,"children":3768},{},[3769],{"type":70,"value":3770},"77.9 GB",{"type":64,"tag":1964,"props":3772,"children":3773},{},[3774,3779],{"type":64,"tag":1986,"props":3775,"children":3776},{},[3777],{"type":70,"value":3778},"Episodes",{"type":64,"tag":1986,"props":3780,"children":3781},{},[3782],{"type":70,"value":3783},"1000 mimic + 1000 Cosmos-augmented (table + wrist cams, depth + seg + normals)",{"type":64,"tag":1964,"props":3785,"children":3786},{},[3787,3792],{"type":64,"tag":1986,"props":3788,"children":3789},{},[3790],{"type":70,"value":3791},"Trick",{"type":64,"tag":1986,"props":3793,"children":3794},{},[3795],{"type":70,"value":3796},"10 human teleops → MimicGen 1k → Cosmos Transfer1 photoreal domain randomization",{"type":64,"tag":1964,"props":3798,"children":3799},{},[3800,3804],{"type":64,"tag":1986,"props":3801,"children":3802},{},[3803],{"type":70,"value":2628},{"type":64,"tag":1986,"props":3805,"children":3806},{},[3807],{"type":70,"value":3712},{"type":64,"tag":1964,"props":3809,"children":3810},{},[3811,3815],{"type":64,"tag":1986,"props":3812,"children":3813},{},[3814],{"type":70,"value":3203},{"type":64,"tag":1986,"props":3816,"children":3817},{},[3818,3824],{"type":64,"tag":206,"props":3819,"children":3822},{"href":3820,"rel":3821},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2503.14492",[209],[3823],{"type":70,"value":3820},{"type":70,"value":3825}," (Cosmos-Transfer1)",{"type":64,"tag":1964,"props":3827,"children":3828},{},[3829,3833],{"type":64,"tag":1986,"props":3830,"children":3831},{},[3832],{"type":70,"value":2717},{"type":64,"tag":1986,"props":3834,"children":3835},{},[3836,3837,3842],{"type":70,"value":3224},{"type":64,"tag":86,"props":3838,"children":3840},{"className":3839},[],[3841],{"type":70,"value":3230},{"type":70,"value":3843}," (legacy Transfer1 workflow), Isaac Sim \u002F Isaac Lab (replay scripts ship in the dataset repo).",{"type":64,"tag":319,"props":3845,"children":3847},{"id":3846},"physicalai-robotics-manipulation-kitchen-demos",[3848],{"type":70,"value":2294},{"type":64,"tag":80,"props":3850,"children":3851},{},[3852],{"type":70,"value":3853},"Massive human-teleop dataset on Franka + Omron mobile base.",{"type":64,"tag":1956,"props":3855,"children":3856},{},[3857,3871],{"type":64,"tag":1960,"props":3858,"children":3859},{},[3860],{"type":64,"tag":1964,"props":3861,"children":3862},{},[3863,3867],{"type":64,"tag":1968,"props":3864,"children":3865},{},[3866],{"type":70,"value":2533},{"type":64,"tag":1968,"props":3868,"children":3869},{},[3870],{"type":70,"value":2538},{"type":64,"tag":1979,"props":3872,"children":3873},{},[3874,3890,3902,3930,3942,3960],{"type":64,"tag":1964,"props":3875,"children":3876},{},[3877,3881],{"type":64,"tag":1986,"props":3878,"children":3879},{},[3880],{"type":70,"value":2549},{"type":64,"tag":1986,"props":3882,"children":3883},{},[3884],{"type":64,"tag":206,"props":3885,"children":3888},{"href":3886,"rel":3887},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-Manipulation-Kitchen-Demos",[209],[3889],{"type":70,"value":3886},{"type":64,"tag":1964,"props":3891,"children":3892},{},[3893,3897],{"type":64,"tag":1986,"props":3894,"children":3895},{},[3896],{"type":70,"value":2566},{"type":64,"tag":1986,"props":3898,"children":3899},{},[3900],{"type":70,"value":3901},"600 h, 55k trajectories, 316 tasks",{"type":64,"tag":1964,"props":3903,"children":3904},{},[3905,3909],{"type":64,"tag":1986,"props":3906,"children":3907},{},[3908],{"type":70,"value":2615},{"type":64,"tag":1986,"props":3910,"children":3911},{},[3912,3914,3920,3922,3928],{"type":70,"value":3913},"LeRobot v2.x with MuJoCo ",{"type":64,"tag":86,"props":3915,"children":3917},{"className":3916},[],[3918],{"type":70,"value":3919},"extras\u002F",{"type":70,"value":3921}," (",{"type":64,"tag":86,"props":3923,"children":3925},{"className":3924},[],[3926],{"type":70,"value":3927},"model.xml.gz",{"type":70,"value":3929}," + raw states)",{"type":64,"tag":1964,"props":3931,"children":3932},{},[3933,3937],{"type":64,"tag":1986,"props":3934,"children":3935},{},[3936],{"type":70,"value":3143},{"type":64,"tag":1986,"props":3938,"children":3939},{},[3940],{"type":70,"value":3941},"left + right agentview + eye-in-hand",{"type":64,"tag":1964,"props":3943,"children":3944},{},[3945,3949],{"type":64,"tag":1986,"props":3946,"children":3947},{},[3948],{"type":70,"value":3402},{"type":64,"tag":1986,"props":3950,"children":3951},{},[3952,3958],{"type":64,"tag":86,"props":3953,"children":3955},{"className":3954},[],[3956],{"type":70,"value":3957},"pretrain\u002Fatomic\u002F...",{"type":70,"value":3959}," × 100 traj\u002Ftask (Open*, Close*, PickPlace*, Adjust*, Coffee*, NavigateKitchen, …)",{"type":64,"tag":1964,"props":3961,"children":3962},{},[3963,3967],{"type":64,"tag":1986,"props":3964,"children":3965},{},[3966],{"type":70,"value":2717},{"type":64,"tag":1986,"props":3968,"children":3969},{},[3970],{"type":70,"value":3971},"Pair with the MJCF assets dataset below for replay in MuJoCo.",{"type":64,"tag":319,"props":3973,"children":3975},{"id":3974},"physicalai-robotics-manipulation-objects-kitchen-mjcf",[3976],{"type":70,"value":2313},{"type":64,"tag":80,"props":3978,"children":3979},{},[3980],{"type":70,"value":3981},"The MuJoCo XML assets that back the Kitchen-Demos environment.",{"type":64,"tag":1956,"props":3983,"children":3984},{},[3985,3999],{"type":64,"tag":1960,"props":3986,"children":3987},{},[3988],{"type":64,"tag":1964,"props":3989,"children":3990},{},[3991,3995],{"type":64,"tag":1968,"props":3992,"children":3993},{},[3994],{"type":70,"value":2533},{"type":64,"tag":1968,"props":3996,"children":3997},{},[3998],{"type":70,"value":2538},{"type":64,"tag":1979,"props":4000,"children":4001},{},[4002,4018,4030,4043,4063],{"type":64,"tag":1964,"props":4003,"children":4004},{},[4005,4009],{"type":64,"tag":1986,"props":4006,"children":4007},{},[4008],{"type":70,"value":2549},{"type":64,"tag":1986,"props":4010,"children":4011},{},[4012],{"type":64,"tag":206,"props":4013,"children":4016},{"href":4014,"rel":4015},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-Manipulation-Objects-Kitchen-MJCF",[209],[4017],{"type":70,"value":4014},{"type":64,"tag":1964,"props":4019,"children":4020},{},[4021,4025],{"type":64,"tag":1986,"props":4022,"children":4023},{},[4024],{"type":70,"value":2566},{"type":64,"tag":1986,"props":4026,"children":4027},{},[4028],{"type":70,"value":4029},"1.32 GB",{"type":64,"tag":1964,"props":4031,"children":4032},{},[4033,4038],{"type":64,"tag":1986,"props":4034,"children":4035},{},[4036],{"type":70,"value":4037},"Categories",{"type":64,"tag":1986,"props":4039,"children":4040},{},[4041],{"type":70,"value":4042},"Objects (~58 categories from kettle to whisk) + Fixtures (12 — blender, coffee machine, dishwasher, electric kettle, fridge, microwave, oven, stand mixer, stove, toaster, toaster oven, cabinet panel)",{"type":64,"tag":1964,"props":4044,"children":4045},{},[4046,4050],{"type":64,"tag":1986,"props":4047,"children":4048},{},[4049],{"type":70,"value":2615},{"type":64,"tag":1986,"props":4051,"children":4052},{},[4053,4055,4061],{"type":70,"value":4054},"Per-model ",{"type":64,"tag":86,"props":4056,"children":4058},{"className":4057},[],[4059],{"type":70,"value":4060},"model.xml",{"type":70,"value":4062}," + visual \u002F collision OBJ + textures, zipped per category",{"type":64,"tag":1964,"props":4064,"children":4065},{},[4066,4070],{"type":64,"tag":1986,"props":4067,"children":4068},{},[4069],{"type":70,"value":2717},{"type":64,"tag":1986,"props":4071,"children":4072},{},[4073,4075,4081],{"type":70,"value":4074},"MuJoCo replay of ",{"type":64,"tag":86,"props":4076,"children":4078},{"className":4077},[],[4079],{"type":70,"value":4080},"Manipulation-Kitchen-Demos",{"type":70,"value":621},{"type":64,"tag":73,"props":4083,"children":4085},{"id":4084},"robotics-gr00t",[4086],{"type":70,"value":711},{"type":64,"tag":80,"props":4088,"children":4089},{},[4090],{"type":70,"value":4091},"GR00T = NVIDIA's generalist humanoid foundation model line. Most data is\nsim-generated for post-training; eval \u002F real-robot supplements are\nsmall.",{"type":64,"tag":319,"props":4093,"children":4095},{"id":4094},"physicalai-robotics-gr00t-x-embodiment-sim",[4096],{"type":70,"value":2100},{"type":64,"tag":80,"props":4098,"children":4099},{},[4100],{"type":70,"value":4101},"Largest GR00T post-training corpus.",{"type":64,"tag":1956,"props":4103,"children":4104},{},[4105,4119],{"type":64,"tag":1960,"props":4106,"children":4107},{},[4108],{"type":64,"tag":1964,"props":4109,"children":4110},{},[4111,4115],{"type":64,"tag":1968,"props":4112,"children":4113},{},[4114],{"type":70,"value":2533},{"type":64,"tag":1968,"props":4116,"children":4117},{},[4118],{"type":70,"value":2538},{"type":64,"tag":1979,"props":4120,"children":4121},{},[4122,4138,4153,4166,4218],{"type":64,"tag":1964,"props":4123,"children":4124},{},[4125,4129],{"type":64,"tag":1986,"props":4126,"children":4127},{},[4128],{"type":70,"value":2549},{"type":64,"tag":1986,"props":4130,"children":4131},{},[4132],{"type":64,"tag":206,"props":4133,"children":4136},{"href":4134,"rel":4135},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-GR00T-X-Embodiment-Sim",[209],[4137],{"type":70,"value":4134},{"type":64,"tag":1964,"props":4139,"children":4140},{},[4141,4145],{"type":64,"tag":1986,"props":4142,"children":4143},{},[4144],{"type":70,"value":2566},{"type":64,"tag":1986,"props":4146,"children":4147},{},[4148],{"type":64,"tag":129,"props":4149,"children":4150},{},[4151],{"type":70,"value":4152},"1.91 TB",{"type":64,"tag":1964,"props":4154,"children":4155},{},[4156,4161],{"type":64,"tag":1986,"props":4157,"children":4158},{},[4159],{"type":70,"value":4160},"Composition",{"type":64,"tag":1986,"props":4162,"children":4163},{},[4164],{"type":70,"value":4165},"9k cross-embodied bimanual (Panda + GR1) + 240k humanoid GR1 tabletop + 24k downsampled + 72k single-Panda RoboCasa + 102 Unitree G1 loco-manipulation = ~345k trajectories",{"type":64,"tag":1964,"props":4167,"children":4168},{},[4169,4174],{"type":64,"tag":1986,"props":4170,"children":4171},{},[4172],{"type":70,"value":4173},"Used by",{"type":64,"tag":1986,"props":4175,"children":4176},{},[4177,4183,4184,4190,4191,4197,4198,4204,4205,4211,4212],{"type":64,"tag":86,"props":4178,"children":4180},{"className":4179},[],[4181],{"type":70,"value":4182},"nvidia\u002FGR00T-N1.5-3B",{"type":70,"value":101},{"type":64,"tag":86,"props":4185,"children":4187},{"className":4186},[],[4188],{"type":70,"value":4189},"GR00T-N1.6-3B",{"type":70,"value":101},{"type":64,"tag":86,"props":4192,"children":4194},{"className":4193},[],[4195],{"type":70,"value":4196},"GR00T-N1.6-bridge",{"type":70,"value":101},{"type":64,"tag":86,"props":4199,"children":4201},{"className":4200},[],[4202],{"type":70,"value":4203},"GR00T-N1.6-G1-PnPAppleToPlate",{"type":70,"value":101},{"type":64,"tag":86,"props":4206,"children":4208},{"className":4207},[],[4209],{"type":70,"value":4210},"GR00T-N1.6-DROID",{"type":70,"value":101},{"type":64,"tag":86,"props":4213,"children":4215},{"className":4214},[],[4216],{"type":70,"value":4217},"GR00T-N1.6-fractal",{"type":64,"tag":1964,"props":4219,"children":4220},{},[4221,4226],{"type":64,"tag":1986,"props":4222,"children":4223},{},[4224],{"type":70,"value":4225},"Download tip",{"type":64,"tag":1986,"props":4227,"children":4228},{},[4229,4231,4237],{"type":70,"value":4230},"Always pass ",{"type":64,"tag":86,"props":4232,"children":4234},{"className":4233},[],[4235],{"type":70,"value":4236},"--include \"\u003Ctask>\u002F**\"",{"type":70,"value":4238}," — full clone is 1.91 TB",{"type":64,"tag":319,"props":4240,"children":4242},{"id":4241},"physicalai-gr00t-tuned-tasks",[4243],{"type":70,"value":2119},{"type":64,"tag":80,"props":4245,"children":4246},{},[4247],{"type":70,"value":4248},"Two industrial post-training task families.",{"type":64,"tag":1956,"props":4250,"children":4251},{},[4252,4266],{"type":64,"tag":1960,"props":4253,"children":4254},{},[4255],{"type":64,"tag":1964,"props":4256,"children":4257},{},[4258,4262],{"type":64,"tag":1968,"props":4259,"children":4260},{},[4261],{"type":70,"value":2533},{"type":64,"tag":1968,"props":4263,"children":4264},{},[4265],{"type":70,"value":2538},{"type":64,"tag":1979,"props":4267,"children":4268},{},[4269,4285,4297,4309,4321,4332],{"type":64,"tag":1964,"props":4270,"children":4271},{},[4272,4276],{"type":64,"tag":1986,"props":4273,"children":4274},{},[4275],{"type":70,"value":2549},{"type":64,"tag":1986,"props":4277,"children":4278},{},[4279],{"type":64,"tag":206,"props":4280,"children":4283},{"href":4281,"rel":4282},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-GR00T-Tuned-Tasks",[209],[4284],{"type":70,"value":4281},{"type":64,"tag":1964,"props":4286,"children":4287},{},[4288,4292],{"type":64,"tag":1986,"props":4289,"children":4290},{},[4291],{"type":70,"value":2566},{"type":64,"tag":1986,"props":4293,"children":4294},{},[4295],{"type":70,"value":4296},"26.5 GB",{"type":64,"tag":1964,"props":4298,"children":4299},{},[4300,4304],{"type":64,"tag":1986,"props":4301,"children":4302},{},[4303],{"type":70,"value":3402},{"type":64,"tag":1986,"props":4305,"children":4306},{},[4307],{"type":70,"value":4308},"Exhaust-Pipe-Sorting (1000), Nut-Pouring (1000)",{"type":64,"tag":1964,"props":4310,"children":4311},{},[4312,4316],{"type":64,"tag":1986,"props":4313,"children":4314},{},[4315],{"type":70,"value":2615},{"type":64,"tag":1986,"props":4317,"children":4318},{},[4319],{"type":70,"value":4320},"HDF5 + GR00T-LeRobot, 256² first-person RGB, 26-DoF state\u002Faction, 20 Hz",{"type":64,"tag":1964,"props":4322,"children":4323},{},[4324,4328],{"type":64,"tag":1986,"props":4325,"children":4326},{},[4327],{"type":70,"value":2628},{"type":64,"tag":1986,"props":4329,"children":4330},{},[4331],{"type":70,"value":3712},{"type":64,"tag":1964,"props":4333,"children":4334},{},[4335,4340],{"type":64,"tag":1986,"props":4336,"children":4337},{},[4338],{"type":70,"value":4339},"Models",{"type":64,"tag":1986,"props":4341,"children":4342},{},[4343,4349,4350],{"type":64,"tag":86,"props":4344,"children":4346},{"className":4345},[],[4347],{"type":70,"value":4348},"nvidia\u002FGR00T-N1-2B-tuned-Nut-Pouring-task",{"type":70,"value":101},{"type":64,"tag":86,"props":4351,"children":4353},{"className":4352},[],[4354],{"type":70,"value":4355},"…-Exhaust-Pipe-Sorting-task",{"type":64,"tag":319,"props":4357,"children":4359},{"id":4358},"physicalai-robotics-gr00t-teleop-g1",[4360],{"type":70,"value":2162},{"type":64,"tag":80,"props":4362,"children":4363},{},[4364],{"type":70,"value":4365},"Real-robot Unitree G1 fruit pick-and-place.",{"type":64,"tag":1956,"props":4367,"children":4368},{},[4369,4383],{"type":64,"tag":1960,"props":4370,"children":4371},{},[4372],{"type":64,"tag":1964,"props":4373,"children":4374},{},[4375,4379],{"type":64,"tag":1968,"props":4376,"children":4377},{},[4378],{"type":70,"value":2533},{"type":64,"tag":1968,"props":4380,"children":4381},{},[4382],{"type":70,"value":2538},{"type":64,"tag":1979,"props":4384,"children":4385},{},[4386,4402,4414,4426,4438,4450,4461],{"type":64,"tag":1964,"props":4387,"children":4388},{},[4389,4393],{"type":64,"tag":1986,"props":4390,"children":4391},{},[4392],{"type":70,"value":2549},{"type":64,"tag":1986,"props":4394,"children":4395},{},[4396],{"type":64,"tag":206,"props":4397,"children":4400},{"href":4398,"rel":4399},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-GR00T-Teleop-G1",[209],[4401],{"type":70,"value":4398},{"type":64,"tag":1964,"props":4403,"children":4404},{},[4405,4409],{"type":64,"tag":1986,"props":4406,"children":4407},{},[4408],{"type":70,"value":2566},{"type":64,"tag":1986,"props":4410,"children":4411},{},[4412],{"type":70,"value":4413},"534 MB",{"type":64,"tag":1964,"props":4415,"children":4416},{},[4417,4421],{"type":64,"tag":1986,"props":4418,"children":4419},{},[4420],{"type":70,"value":3415},{"type":64,"tag":1986,"props":4422,"children":4423},{},[4424],{"type":70,"value":4425},"1000 real teleop, Unitree G1 upper body + Tri-finger hands + RealSense",{"type":64,"tag":1964,"props":4427,"children":4428},{},[4429,4433],{"type":64,"tag":1986,"props":4430,"children":4431},{},[4432],{"type":70,"value":3402},{"type":64,"tag":1986,"props":4434,"children":4435},{},[4436],{"type":70,"value":4437},"Pick {apple, pear, grapes, starfruit} → basket",{"type":64,"tag":1964,"props":4439,"children":4440},{},[4441,4445],{"type":64,"tag":1986,"props":4442,"children":4443},{},[4444],{"type":70,"value":2615},{"type":64,"tag":1986,"props":4446,"children":4447},{},[4448],{"type":70,"value":4449},"MP4 + HDF5",{"type":64,"tag":1964,"props":4451,"children":4452},{},[4453,4457],{"type":64,"tag":1986,"props":4454,"children":4455},{},[4456],{"type":70,"value":2628},{"type":64,"tag":1986,"props":4458,"children":4459},{},[4460],{"type":70,"value":3712},{"type":64,"tag":1964,"props":4462,"children":4463},{},[4464,4468],{"type":64,"tag":1986,"props":4465,"children":4466},{},[4467],{"type":70,"value":2717},{"type":64,"tag":1986,"props":4469,"children":4470},{},[4471,4477,4479,4485],{"type":64,"tag":86,"props":4472,"children":4474},{"className":4473},[],[4475],{"type":70,"value":4476},"Isaac-GR00T",{"type":70,"value":4478}," finetune docs (",{"type":64,"tag":86,"props":4480,"children":4482},{"className":4481},[],[4483],{"type":70,"value":4484},"getting_started\u002F3_0_new_embodiment_finetuning.md",{"type":70,"value":2019},{"type":64,"tag":319,"props":4487,"children":4489},{"id":4488},"physicalai-robotics-gr00t-teleop-sim",[4490],{"type":70,"value":2181},{"type":64,"tag":80,"props":4492,"children":4493},{},[4494],{"type":70,"value":4495},"Simulated GR1 tabletop teleop.",{"type":64,"tag":1956,"props":4497,"children":4498},{},[4499,4513],{"type":64,"tag":1960,"props":4500,"children":4501},{},[4502],{"type":64,"tag":1964,"props":4503,"children":4504},{},[4505,4509],{"type":64,"tag":1968,"props":4506,"children":4507},{},[4508],{"type":70,"value":2533},{"type":64,"tag":1968,"props":4510,"children":4511},{},[4512],{"type":70,"value":2538},{"type":64,"tag":1979,"props":4514,"children":4515},{},[4516,4532,4544,4556,4573],{"type":64,"tag":1964,"props":4517,"children":4518},{},[4519,4523],{"type":64,"tag":1986,"props":4520,"children":4521},{},[4522],{"type":70,"value":2549},{"type":64,"tag":1986,"props":4524,"children":4525},{},[4526],{"type":64,"tag":206,"props":4527,"children":4530},{"href":4528,"rel":4529},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-GR00T-Teleop-Sim",[209],[4531],{"type":70,"value":4528},{"type":64,"tag":1964,"props":4533,"children":4534},{},[4535,4539],{"type":64,"tag":1986,"props":4536,"children":4537},{},[4538],{"type":70,"value":2566},{"type":64,"tag":1986,"props":4540,"children":4541},{},[4542],{"type":70,"value":4543},"55.4 GB (39 GB LeRobot + 14 GB HDF5)",{"type":64,"tag":1964,"props":4545,"children":4546},{},[4547,4551],{"type":64,"tag":1986,"props":4548,"children":4549},{},[4550],{"type":70,"value":3415},{"type":64,"tag":1986,"props":4552,"children":4553},{},[4554],{"type":70,"value":4555},"24 tasks × 1000 each",{"type":64,"tag":1964,"props":4557,"children":4558},{},[4559,4563],{"type":64,"tag":1986,"props":4560,"children":4561},{},[4562],{"type":70,"value":2628},{"type":64,"tag":1986,"props":4564,"children":4565},{},[4566,4571],{"type":64,"tag":129,"props":4567,"children":4568},{},[4569],{"type":70,"value":4570},"CC-BY-NC-4.0",{"type":70,"value":4572}," (non-commercial) — different from G1 above",{"type":64,"tag":1964,"props":4574,"children":4575},{},[4576,4580],{"type":64,"tag":1986,"props":4577,"children":4578},{},[4579],{"type":70,"value":2615},{"type":64,"tag":1986,"props":4581,"children":4582},{},[4583],{"type":70,"value":4584},"HDF5 + LeRobot",{"type":64,"tag":319,"props":4586,"children":4588},{"id":4587},"physicalai-robotics-gr00t-teleop-gr1",[4589],{"type":70,"value":2200},{"type":64,"tag":80,"props":4591,"children":4592},{},[4593],{"type":70,"value":4594},"DreamDojo pretraining corpus — large-scale human egocentric video.",{"type":64,"tag":1956,"props":4596,"children":4597},{},[4598,4612],{"type":64,"tag":1960,"props":4599,"children":4600},{},[4601],{"type":64,"tag":1964,"props":4602,"children":4603},{},[4604,4608],{"type":64,"tag":1968,"props":4605,"children":4606},{},[4607],{"type":70,"value":2533},{"type":64,"tag":1968,"props":4609,"children":4610},{},[4611],{"type":70,"value":2538},{"type":64,"tag":1979,"props":4613,"children":4614},{},[4615,4631,4643,4656,4681],{"type":64,"tag":1964,"props":4616,"children":4617},{},[4618,4622],{"type":64,"tag":1986,"props":4619,"children":4620},{},[4621],{"type":70,"value":2549},{"type":64,"tag":1986,"props":4623,"children":4624},{},[4625],{"type":64,"tag":206,"props":4626,"children":4629},{"href":4627,"rel":4628},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-GR00T-Teleop-GR1",[209],[4630],{"type":70,"value":4627},{"type":64,"tag":1964,"props":4632,"children":4633},{},[4634,4638],{"type":64,"tag":1986,"props":4635,"children":4636},{},[4637],{"type":70,"value":2566},{"type":64,"tag":1986,"props":4639,"children":4640},{},[4641],{"type":70,"value":4642},"74.3 GB",{"type":64,"tag":1964,"props":4644,"children":4645},{},[4646,4651],{"type":64,"tag":1986,"props":4647,"children":4648},{},[4649],{"type":70,"value":4650},"Coverage",{"type":64,"tag":1986,"props":4652,"children":4653},{},[4654],{"type":70,"value":4655},"44k hours of human egocentric data (per project page)",{"type":64,"tag":1964,"props":4657,"children":4658},{},[4659,4664],{"type":64,"tag":1986,"props":4660,"children":4661},{},[4662],{"type":70,"value":4663},"Project",{"type":64,"tag":1986,"props":4665,"children":4666},{},[4667,4673,4675],{"type":64,"tag":206,"props":4668,"children":4671},{"href":4669,"rel":4670},"https:\u002F\u002Fdreamdojo-world.github.io\u002F",[209],[4672],{"type":70,"value":4669},{"type":70,"value":4674}," + ",{"type":64,"tag":206,"props":4676,"children":4679},{"href":4677,"rel":4678},"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FDreamDojo",[209],[4680],{"type":70,"value":4677},{"type":64,"tag":1964,"props":4682,"children":4683},{},[4684,4688],{"type":64,"tag":1986,"props":4685,"children":4686},{},[4687],{"type":70,"value":3203},{"type":64,"tag":1986,"props":4689,"children":4690},{},[4691],{"type":64,"tag":206,"props":4692,"children":4695},{"href":4693,"rel":4694},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2602.06949",[209],[4696],{"type":70,"value":4693},{"type":64,"tag":319,"props":4698,"children":4700},{"id":4699},"physicalai-robotics-gr00t-gr1",[4701],{"type":70,"value":2145},{"type":64,"tag":80,"props":4703,"children":4704},{},[4705],{"type":70,"value":4706},"Lab-recorded GR1-T2 third-person video.",{"type":64,"tag":1956,"props":4708,"children":4709},{},[4710,4724],{"type":64,"tag":1960,"props":4711,"children":4712},{},[4713],{"type":64,"tag":1964,"props":4714,"children":4715},{},[4716,4720],{"type":64,"tag":1968,"props":4717,"children":4718},{},[4719],{"type":70,"value":2533},{"type":64,"tag":1968,"props":4721,"children":4722},{},[4723],{"type":70,"value":2538},{"type":64,"tag":1979,"props":4725,"children":4726},{},[4727,4743,4755,4768,4781],{"type":64,"tag":1964,"props":4728,"children":4729},{},[4730,4734],{"type":64,"tag":1986,"props":4731,"children":4732},{},[4733],{"type":70,"value":2549},{"type":64,"tag":1986,"props":4735,"children":4736},{},[4737],{"type":64,"tag":206,"props":4738,"children":4741},{"href":4739,"rel":4740},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-GR00T-GR1",[209],[4742],{"type":70,"value":4739},{"type":64,"tag":1964,"props":4744,"children":4745},{},[4746,4750],{"type":64,"tag":1986,"props":4747,"children":4748},{},[4749],{"type":70,"value":2566},{"type":64,"tag":1986,"props":4751,"children":4752},{},[4753],{"type":70,"value":4754},"142 MB",{"type":64,"tag":1964,"props":4756,"children":4757},{},[4758,4763],{"type":64,"tag":1986,"props":4759,"children":4760},{},[4761],{"type":70,"value":4762},"Records",{"type":64,"tag":1986,"props":4764,"children":4765},{},[4766],{"type":70,"value":4767},"92 MP4 videos (Fourier GR1-T2)",{"type":64,"tag":1964,"props":4769,"children":4770},{},[4771,4776],{"type":64,"tag":1986,"props":4772,"children":4773},{},[4774],{"type":70,"value":4775},"Use",{"type":64,"tag":1986,"props":4777,"children":4778},{},[4779],{"type":70,"value":4780},"DreamGen training reference",{"type":64,"tag":1964,"props":4782,"children":4783},{},[4784,4788],{"type":64,"tag":1986,"props":4785,"children":4786},{},[4787],{"type":70,"value":2628},{"type":64,"tag":1986,"props":4789,"children":4790},{},[4791],{"type":70,"value":3712},{"type":64,"tag":319,"props":4793,"children":4795},{"id":4794},"physicalai-robotics-gr00t-eval",[4796],{"type":70,"value":2138},{"type":64,"tag":80,"props":4798,"children":4799},{},[4800],{"type":70,"value":4801},"GR00T eval initial-state frames.",{"type":64,"tag":1956,"props":4803,"children":4804},{},[4805,4819],{"type":64,"tag":1960,"props":4806,"children":4807},{},[4808],{"type":64,"tag":1964,"props":4809,"children":4810},{},[4811,4815],{"type":64,"tag":1968,"props":4812,"children":4813},{},[4814],{"type":70,"value":2533},{"type":64,"tag":1968,"props":4816,"children":4817},{},[4818],{"type":70,"value":2538},{"type":64,"tag":1979,"props":4820,"children":4821},{},[4822,4838,4850,4862],{"type":64,"tag":1964,"props":4823,"children":4824},{},[4825,4829],{"type":64,"tag":1986,"props":4826,"children":4827},{},[4828],{"type":70,"value":2549},{"type":64,"tag":1986,"props":4830,"children":4831},{},[4832],{"type":64,"tag":206,"props":4833,"children":4836},{"href":4834,"rel":4835},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-GR00T-Eval",[209],[4837],{"type":70,"value":4834},{"type":64,"tag":1964,"props":4839,"children":4840},{},[4841,4845],{"type":64,"tag":1986,"props":4842,"children":4843},{},[4844],{"type":70,"value":2566},{"type":64,"tag":1986,"props":4846,"children":4847},{},[4848],{"type":70,"value":4849},"180 MB",{"type":64,"tag":1964,"props":4851,"children":4852},{},[4853,4857],{"type":64,"tag":1986,"props":4854,"children":4855},{},[4856],{"type":70,"value":4762},{"type":64,"tag":1986,"props":4858,"children":4859},{},[4860],{"type":70,"value":4861},"123 PNG frames (GR1-T2 robot's first-person view) + per-frame TXT",{"type":64,"tag":1964,"props":4863,"children":4864},{},[4865,4869],{"type":64,"tag":1986,"props":4866,"children":4867},{},[4868],{"type":70,"value":2628},{"type":64,"tag":1986,"props":4870,"children":4871},{},[4872],{"type":70,"value":3176},{"type":64,"tag":73,"props":4874,"children":4876},{"id":4875},"robotics-mindmap",[4877],{"type":70,"value":722},{"type":64,"tag":80,"props":4879,"children":4880},{},[4881,4883,4889,4891,4897,4899,4904,4906,4917],{"type":70,"value":4882},"Spatial-memory benchmark from ",{"type":64,"tag":86,"props":4884,"children":4886},{"className":4885},[],[4887],{"type":70,"value":4888},"nvidia-isaac\u002Fnvblox_mindmap",{"type":70,"value":4890},". Each\ndataset is one task with the same multimodal layout (RGB-D + camera\nintrinsics\u002Fposes + nvblox vertex features in ",{"type":64,"tag":86,"props":4892,"children":4894},{"className":4893},[],[4895],{"type":70,"value":4896},".zst",{"type":70,"value":4898}," + robot state).\n",{"type":64,"tag":129,"props":4900,"children":4901},{},[4902],{"type":70,"value":4903},"All four are CC-BY-NC-4.0",{"type":70,"value":4905}," (research only). Models trained:\n",{"type":64,"tag":206,"props":4907,"children":4910},{"href":4908,"rel":4909},"https:\u002F\u002Fhuggingface.co\u002Fnvidia\u002FPhysicalAI-Robotics-mindmap-Checkpoints",[209],[4911],{"type":64,"tag":86,"props":4912,"children":4914},{"className":4913},[],[4915],{"type":70,"value":4916},"nvidia\u002FPhysicalAI-Robotics-mindmap-Checkpoints",{"type":70,"value":621},{"type":64,"tag":1956,"props":4919,"children":4920},{},[4921,4957],{"type":64,"tag":1960,"props":4922,"children":4923},{},[4924],{"type":64,"tag":1964,"props":4925,"children":4926},{},[4927,4932,4937,4942,4947,4952],{"type":64,"tag":1968,"props":4928,"children":4929},{},[4930],{"type":70,"value":4931},"Dataset",{"type":64,"tag":1968,"props":4933,"children":4934},{},[4935],{"type":70,"value":4936},"Robot",{"type":64,"tag":1968,"props":4938,"children":4939},{},[4940],{"type":70,"value":4941},"Teleop tool",{"type":64,"tag":1968,"props":4943,"children":4944},{},[4945],{"type":70,"value":4946},"Demos provided",{"type":64,"tag":1968,"props":4948,"children":4949},{},[4950],{"type":70,"value":4951},"Total mimic-generated",{"type":64,"tag":1968,"props":4953,"children":4954},{},[4955],{"type":70,"value":4956},"Storage",{"type":64,"tag":1979,"props":4958,"children":4959},{},[4960,5010,5049,5090],{"type":64,"tag":1964,"props":4961,"children":4962},{},[4963,4977,4982,4987,5000,5005],{"type":64,"tag":1986,"props":4964,"children":4965},{},[4966],{"type":64,"tag":206,"props":4967,"children":4970},{"href":4968,"rel":4969},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-mindmap-GR1-Stick-in-Bin",[209],[4971],{"type":64,"tag":86,"props":4972,"children":4974},{"className":4973},[],[4975],{"type":70,"value":4976},"PhysicalAI-Robotics-mindmap-GR1-Stick-in-Bin",{"type":64,"tag":1986,"props":4978,"children":4979},{},[4980],{"type":70,"value":4981},"Fourier GR1",{"type":64,"tag":1986,"props":4983,"children":4984},{},[4985],{"type":70,"value":4986},"Apple Vision Pro",{"type":64,"tag":1986,"props":4988,"children":4989},{},[4990,4992,4998],{"type":70,"value":4991},"10 (",{"type":64,"tag":86,"props":4993,"children":4995},{"className":4994},[],[4996],{"type":70,"value":4997},"mindmap",{"type":70,"value":4999}," fmt) + HDF5",{"type":64,"tag":1986,"props":5001,"children":5002},{},[5003],{"type":70,"value":5004},"200 (from 20 human teleops)",{"type":64,"tag":1986,"props":5006,"children":5007},{},[5008],{"type":70,"value":5009},"103 GB",{"type":64,"tag":1964,"props":5011,"children":5012},{},[5013,5027,5031,5035,5040,5044],{"type":64,"tag":1986,"props":5014,"children":5015},{},[5016],{"type":64,"tag":206,"props":5017,"children":5020},{"href":5018,"rel":5019},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-mindmap-GR1-Drill-in-Box",[209],[5021],{"type":64,"tag":86,"props":5022,"children":5024},{"className":5023},[],[5025],{"type":70,"value":5026},"PhysicalAI-Robotics-mindmap-GR1-Drill-in-Box",{"type":64,"tag":1986,"props":5028,"children":5029},{},[5030],{"type":70,"value":4981},{"type":64,"tag":1986,"props":5032,"children":5033},{},[5034],{"type":70,"value":4986},{"type":64,"tag":1986,"props":5036,"children":5037},{},[5038],{"type":70,"value":5039},"10 + HDF5",{"type":64,"tag":1986,"props":5041,"children":5042},{},[5043],{"type":70,"value":5004},{"type":64,"tag":1986,"props":5045,"children":5046},{},[5047],{"type":70,"value":5048},"53.7 GB",{"type":64,"tag":1964,"props":5050,"children":5051},{},[5052,5066,5071,5076,5080,5085],{"type":64,"tag":1986,"props":5053,"children":5054},{},[5055],{"type":64,"tag":206,"props":5056,"children":5059},{"href":5057,"rel":5058},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-mindmap-Franka-Cube-Stacking",[209],[5060],{"type":64,"tag":86,"props":5061,"children":5063},{"className":5062},[],[5064],{"type":70,"value":5065},"PhysicalAI-Robotics-mindmap-Franka-Cube-Stacking",{"type":64,"tag":1986,"props":5067,"children":5068},{},[5069],{"type":70,"value":5070},"Franka",{"type":64,"tag":1986,"props":5072,"children":5073},{},[5074],{"type":70,"value":5075},"SpaceMouse",{"type":64,"tag":1986,"props":5077,"children":5078},{},[5079],{"type":70,"value":5039},{"type":64,"tag":1986,"props":5081,"children":5082},{},[5083],{"type":70,"value":5084},"1000 (from 10 human teleops)",{"type":64,"tag":1986,"props":5086,"children":5087},{},[5088],{"type":70,"value":5089},"6.12 GB",{"type":64,"tag":1964,"props":5091,"children":5092},{},[5093,5107,5111,5115,5119,5124],{"type":64,"tag":1986,"props":5094,"children":5095},{},[5096],{"type":64,"tag":206,"props":5097,"children":5100},{"href":5098,"rel":5099},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-mindmap-Franka-Mug-in-Drawer",[209],[5101],{"type":64,"tag":86,"props":5102,"children":5104},{"className":5103},[],[5105],{"type":70,"value":5106},"PhysicalAI-Robotics-mindmap-Franka-Mug-in-Drawer",{"type":64,"tag":1986,"props":5108,"children":5109},{},[5110],{"type":70,"value":5070},{"type":64,"tag":1986,"props":5112,"children":5113},{},[5114],{"type":70,"value":5075},{"type":64,"tag":1986,"props":5116,"children":5117},{},[5118],{"type":70,"value":5039},{"type":64,"tag":1986,"props":5120,"children":5121},{},[5122],{"type":70,"value":5123},"250 (from 15 human teleops)",{"type":64,"tag":1986,"props":5125,"children":5126},{},[5127],{"type":70,"value":5128},"34.5 GB",{"type":64,"tag":197,"props":5130,"children":5131},{},[5132],{"type":64,"tag":80,"props":5133,"children":5134},{},[5135,5137,5144],{"type":70,"value":5136},"Provided datasets ship 10 mindmap-formatted demos for storage reasons;\nregenerate the full 200\u002F1000\u002F250 with the upstream\n",{"type":64,"tag":206,"props":5138,"children":5141},{"href":5139,"rel":5140},"https:\u002F\u002Fnvidia-isaac.github.io\u002Fnvblox_mindmap\u002Fpages\u002Fdata_generation.html",[209],[5142],{"type":70,"value":5143},"data-generation docs",{"type":70,"value":621},{"type":64,"tag":80,"props":5146,"children":5147},{},[5148,5150,5156,5158,5164],{"type":70,"value":5149},"Paper: ",{"type":64,"tag":206,"props":5151,"children":5154},{"href":5152,"rel":5153},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2509.20297",[209],[5155],{"type":70,"value":5152},{"type":70,"value":5157},". Codebase:\n",{"type":64,"tag":206,"props":5159,"children":5162},{"href":5160,"rel":5161},"https:\u002F\u002Fgithub.com\u002Fnvidia-isaac\u002Fnvblox_mindmap",[209],[5163],{"type":70,"value":5160},{"type":70,"value":621},{"type":64,"tag":73,"props":5166,"children":5168},{"id":5167},"robotics-nurec-sim-ready-scenes",[5169],{"type":70,"value":733},{"type":64,"tag":319,"props":5171,"children":5173},{"id":5172},"physicalai-robotics-nurec",[5174],{"type":70,"value":2366},{"type":64,"tag":80,"props":5176,"children":5177},{},[5178],{"type":70,"value":5179},"Indoor 3DGUT scenes for IsaacSim AMR simulation.",{"type":64,"tag":1956,"props":5181,"children":5182},{},[5183,5197],{"type":64,"tag":1960,"props":5184,"children":5185},{},[5186],{"type":64,"tag":1964,"props":5187,"children":5188},{},[5189,5193],{"type":64,"tag":1968,"props":5190,"children":5191},{},[5192],{"type":70,"value":2533},{"type":64,"tag":1968,"props":5194,"children":5195},{},[5196],{"type":70,"value":2538},{"type":64,"tag":1979,"props":5198,"children":5199},{},[5200,5216,5228,5241,5258,5287,5300,5311],{"type":64,"tag":1964,"props":5201,"children":5202},{},[5203,5207],{"type":64,"tag":1986,"props":5204,"children":5205},{},[5206],{"type":70,"value":2549},{"type":64,"tag":1986,"props":5208,"children":5209},{},[5210],{"type":64,"tag":206,"props":5211,"children":5214},{"href":5212,"rel":5213},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-NuRec",[209],[5215],{"type":70,"value":5212},{"type":64,"tag":1964,"props":5217,"children":5218},{},[5219,5223],{"type":64,"tag":1986,"props":5220,"children":5221},{},[5222],{"type":70,"value":2566},{"type":64,"tag":1986,"props":5224,"children":5225},{},[5226],{"type":70,"value":5227},"62.9 GB",{"type":64,"tag":1964,"props":5229,"children":5230},{},[5231,5236],{"type":64,"tag":1986,"props":5232,"children":5233},{},[5234],{"type":70,"value":5235},"Scenes",{"type":64,"tag":1986,"props":5237,"children":5238},{},[5239],{"type":70,"value":5240},"Nova-Carter (galileo, cafe, wormhole) — stereo, with mesh + occupancy; Zurich offices (lounge, fourth-floor iphone) — mono, no mesh; Endeavor hand-held (andoria, livingroom, wormhole) — stereo with mesh",{"type":64,"tag":1964,"props":5242,"children":5243},{},[5244,5248],{"type":64,"tag":1986,"props":5245,"children":5246},{},[5247],{"type":70,"value":2615},{"type":64,"tag":1986,"props":5249,"children":5250},{},[5251,5253],{"type":70,"value":5252},"USDZ (3DGUT + mesh + occupancy) loadable in ",{"type":64,"tag":129,"props":5254,"children":5255},{},[5256],{"type":70,"value":5257},"Isaac Sim 5.1",{"type":64,"tag":1964,"props":5259,"children":5260},{},[5261,5266],{"type":64,"tag":1986,"props":5262,"children":5263},{},[5264],{"type":70,"value":5265},"Workflows",{"type":64,"tag":1986,"props":5267,"children":5268},{},[5269,5276,5278,5285],{"type":64,"tag":206,"props":5270,"children":5273},{"href":5271,"rel":5272},"https:\u002F\u002Fdocs.nvidia.com\u002Fnurec\u002Frobotics\u002Fneural_reconstruction_stereo.html",[209],[5274],{"type":70,"value":5275},"Stereo NuRec (Isaac ROS + cuSFM + FoundationStereo + nvblox + 3DGURT)",{"type":70,"value":5277}," for Carter; ",{"type":64,"tag":206,"props":5279,"children":5282},{"href":5280,"rel":5281},"https:\u002F\u002Fdocs.nvidia.com\u002Fnurec\u002Frobotics\u002Fneural_reconstruction_mono.html",[209],[5283],{"type":70,"value":5284},"Mono NuRec (COLMAP + 3DGURT)",{"type":70,"value":5286}," for Zurich",{"type":64,"tag":1964,"props":5288,"children":5289},{},[5290,5295],{"type":64,"tag":1986,"props":5291,"children":5292},{},[5293],{"type":70,"value":5294},"Gating",{"type":64,"tag":1986,"props":5296,"children":5297},{},[5298],{"type":70,"value":5299},"Contact-info gate (no separate license)",{"type":64,"tag":1964,"props":5301,"children":5302},{},[5303,5307],{"type":64,"tag":1986,"props":5304,"children":5305},{},[5306],{"type":70,"value":2628},{"type":64,"tag":1986,"props":5308,"children":5309},{},[5310],{"type":70,"value":3176},{"type":64,"tag":1964,"props":5312,"children":5313},{},[5314,5318],{"type":64,"tag":1986,"props":5315,"children":5316},{},[5317],{"type":70,"value":2717},{"type":64,"tag":1986,"props":5319,"children":5320},{},[5321,5329,5331,5342],{"type":64,"tag":206,"props":5322,"children":5323},{"href":3037},[5324],{"type":64,"tag":86,"props":5325,"children":5327},{"className":5326},[],[5328],{"type":70,"value":3037},{"type":70,"value":5330}," to retrain reconstructions; upstream Isaac Sim 5.1 (no in-repo skill) for AMR simulation; pair with ",{"type":64,"tag":206,"props":5332,"children":5335},{"href":5333,"rel":5334},"https:\u002F\u002Fgithub.com\u002FNVlabs\u002FMobilityGen",[209],[5336],{"type":64,"tag":86,"props":5337,"children":5339},{"className":5338},[],[5340],{"type":70,"value":5341},"MobilityGen",{"type":70,"value":5343}," for AMR data generation.",{"type":64,"tag":319,"props":5345,"children":5347},{"id":5346},"physicalai-nurec-ppisp",[5348],{"type":70,"value":2441},{"type":64,"tag":80,"props":5350,"children":5351},{},[5352],{"type":70,"value":5353},"Photometric-variation benchmark for radiance-field methods.",{"type":64,"tag":1956,"props":5355,"children":5356},{},[5357,5371],{"type":64,"tag":1960,"props":5358,"children":5359},{},[5360],{"type":64,"tag":1964,"props":5361,"children":5362},{},[5363,5367],{"type":64,"tag":1968,"props":5364,"children":5365},{},[5366],{"type":70,"value":2533},{"type":64,"tag":1968,"props":5368,"children":5369},{},[5370],{"type":70,"value":2538},{"type":64,"tag":1979,"props":5372,"children":5373},{},[5374,5390,5402,5423,5436,5447],{"type":64,"tag":1964,"props":5375,"children":5376},{},[5377,5381],{"type":64,"tag":1986,"props":5378,"children":5379},{},[5380],{"type":70,"value":2549},{"type":64,"tag":1986,"props":5382,"children":5383},{},[5384],{"type":64,"tag":206,"props":5385,"children":5388},{"href":5386,"rel":5387},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-NuRec-PPISP",[209],[5389],{"type":70,"value":5386},{"type":64,"tag":1964,"props":5391,"children":5392},{},[5393,5397],{"type":64,"tag":1986,"props":5394,"children":5395},{},[5396],{"type":70,"value":2566},{"type":64,"tag":1986,"props":5398,"children":5399},{},[5400],{"type":70,"value":5401},"15.2 GB (8.1 GB COLMAP + 6.2 GB NCore V4)",{"type":64,"tag":1964,"props":5403,"children":5404},{},[5405,5410],{"type":64,"tag":1986,"props":5406,"children":5407},{},[5408],{"type":70,"value":5409},"Captures",{"type":64,"tag":1986,"props":5411,"children":5412},{},[5413,5415,5421],{"type":70,"value":5414},"4 outdoor scenes × 3 cameras (Nikon Z7, OM-1 II, iPhone 13 Pro) = 8 sequences (~2600 photos), ",{"type":64,"tag":86,"props":5416,"children":5418},{"className":5417},[],[5419],{"type":70,"value":5420},"+\u002F-2 EV",{"type":70,"value":5422}," exposure bracketing",{"type":64,"tag":1964,"props":5424,"children":5425},{},[5426,5431],{"type":64,"tag":1986,"props":5427,"children":5428},{},[5429],{"type":70,"value":5430},"Variants",{"type":64,"tag":1986,"props":5432,"children":5433},{},[5434],{"type":70,"value":5435},"Standard (full bracket) + auto (re-processed with auto-exposure \u002F WB)",{"type":64,"tag":1964,"props":5437,"children":5438},{},[5439,5443],{"type":64,"tag":1986,"props":5440,"children":5441},{},[5442],{"type":70,"value":2628},{"type":64,"tag":1986,"props":5444,"children":5445},{},[5446],{"type":70,"value":3712},{"type":64,"tag":1964,"props":5448,"children":5449},{},[5450,5454],{"type":64,"tag":1986,"props":5451,"children":5452},{},[5453],{"type":70,"value":2717},{"type":64,"tag":1986,"props":5455,"children":5456},{},[5457,5459,5467,5468,5474,5475,5483],{"type":70,"value":5458},"3DGRUT \u002F GSplat benchmarking via ",{"type":64,"tag":206,"props":5460,"children":5461},{"href":3037},[5462],{"type":64,"tag":86,"props":5463,"children":5465},{"className":5464},[],[5466],{"type":70,"value":3037},{"type":70,"value":3921},{"type":64,"tag":86,"props":5469,"children":5471},{"className":5470},[],[5472],{"type":70,"value":5473},"eval-rendering-metrics",{"type":70,"value":3053},{"type":64,"tag":206,"props":5476,"children":5477},{"href":2733},[5478],{"type":64,"tag":86,"props":5479,"children":5481},{"className":5480},[],[5482],{"type":70,"value":2733},{"type":70,"value":621},{"type":64,"tag":73,"props":5485,"children":5487},{"id":5486},"robotics-healthcare",[5488],{"type":70,"value":744},{"type":64,"tag":319,"props":5490,"children":5492},{"id":5491},"physicalai-robotics-open-h-embodiment",[5493],{"type":70,"value":2479},{"type":64,"tag":80,"props":5495,"children":5496},{},[5497],{"type":70,"value":5498},"Surgical \u002F ultrasound robotics multi-embodiment corpus.",{"type":64,"tag":1956,"props":5500,"children":5501},{},[5502,5516],{"type":64,"tag":1960,"props":5503,"children":5504},{},[5505],{"type":64,"tag":1964,"props":5506,"children":5507},{},[5508,5512],{"type":64,"tag":1968,"props":5509,"children":5510},{},[5511],{"type":70,"value":2533},{"type":64,"tag":1968,"props":5513,"children":5514},{},[5515],{"type":70,"value":2538},{"type":64,"tag":1979,"props":5517,"children":5518},{},[5519,5535,5547,5559,5572,5598],{"type":64,"tag":1964,"props":5520,"children":5521},{},[5522,5526],{"type":64,"tag":1986,"props":5523,"children":5524},{},[5525],{"type":70,"value":2549},{"type":64,"tag":1986,"props":5527,"children":5528},{},[5529],{"type":64,"tag":206,"props":5530,"children":5533},{"href":5531,"rel":5532},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-Open-H-Embodiment",[209],[5534],{"type":70,"value":5531},{"type":64,"tag":1964,"props":5536,"children":5537},{},[5538,5542],{"type":64,"tag":1986,"props":5539,"children":5540},{},[5541],{"type":70,"value":2566},{"type":64,"tag":1986,"props":5543,"children":5544},{},[5545],{"type":70,"value":5546},"4.5 TB, 750 h, 120,000 trajectories",{"type":64,"tag":1964,"props":5548,"children":5549},{},[5550,5554],{"type":64,"tag":1986,"props":5551,"children":5552},{},[5553],{"type":70,"value":2615},{"type":64,"tag":1986,"props":5555,"children":5556},{},[5557],{"type":70,"value":5558},"LeRobot v2.1 — MP4 video + Parquet kinematics + JSONL manifests",{"type":64,"tag":1964,"props":5560,"children":5561},{},[5562,5567],{"type":64,"tag":1986,"props":5563,"children":5564},{},[5565],{"type":70,"value":5566},"Contributors",{"type":64,"tag":1986,"props":5568,"children":5569},{},[5570],{"type":70,"value":5571},"30+ orgs (JHU, Stanford, UCSD, UCB, Vanderbilt, TUM, MBZUAI, …)",{"type":64,"tag":1964,"props":5573,"children":5574},{},[5575,5579],{"type":64,"tag":1986,"props":5576,"children":5577},{},[5578],{"type":70,"value":78},{"type":64,"tag":1986,"props":5580,"children":5581},{},[5582,5584,5590,5591,5597],{"type":70,"value":5583},"Healthcare autonomy + world-foundation-model training (used by ",{"type":64,"tag":86,"props":5585,"children":5587},{"className":5586},[],[5588],{"type":70,"value":5589},"nvidia\u002FGR00T-H",{"type":70,"value":1396},{"type":64,"tag":86,"props":5592,"children":5594},{"className":5593},[],[5595],{"type":70,"value":5596},"nvidia\u002FCosmos-H-Surgical-Simulator",{"type":70,"value":2019},{"type":64,"tag":1964,"props":5599,"children":5600},{},[5601,5605],{"type":64,"tag":1986,"props":5602,"children":5603},{},[5604],{"type":70,"value":2628},{"type":64,"tag":1986,"props":5606,"children":5607},{},[5608],{"type":70,"value":3176},{"type":64,"tag":73,"props":5610,"children":5612},{"id":5611},"robotics-grasping",[5613],{"type":70,"value":755},{"type":64,"tag":319,"props":5615,"children":5617},{"id":5616},"physicalai-robotics-graspgen",[5618],{"type":70,"value":2460},{"type":64,"tag":80,"props":5620,"children":5621},{},[5622],{"type":70,"value":5623},"Sim2Real grasping at scale.",{"type":64,"tag":1956,"props":5625,"children":5626},{},[5627,5641],{"type":64,"tag":1960,"props":5628,"children":5629},{},[5630],{"type":64,"tag":1964,"props":5631,"children":5632},{},[5633,5637],{"type":64,"tag":1968,"props":5634,"children":5635},{},[5636],{"type":70,"value":2533},{"type":64,"tag":1968,"props":5638,"children":5639},{},[5640],{"type":70,"value":2538},{"type":64,"tag":1979,"props":5642,"children":5643},{},[5644,5660,5672,5689,5702,5722,5733,5754],{"type":64,"tag":1964,"props":5645,"children":5646},{},[5647,5651],{"type":64,"tag":1986,"props":5648,"children":5649},{},[5650],{"type":70,"value":2549},{"type":64,"tag":1986,"props":5652,"children":5653},{},[5654],{"type":64,"tag":206,"props":5655,"children":5658},{"href":5656,"rel":5657},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-GraspGen",[209],[5659],{"type":70,"value":5656},{"type":64,"tag":1964,"props":5661,"children":5662},{},[5663,5667],{"type":64,"tag":1986,"props":5664,"children":5665},{},[5666],{"type":70,"value":2566},{"type":64,"tag":1986,"props":5668,"children":5669},{},[5670],{"type":70,"value":5671},"21.6 GB",{"type":64,"tag":1964,"props":5673,"children":5674},{},[5675,5679],{"type":64,"tag":1986,"props":5676,"children":5677},{},[5678],{"type":70,"value":4650},{"type":64,"tag":1986,"props":5680,"children":5681},{},[5682,5687],{"type":64,"tag":129,"props":5683,"children":5684},{},[5685],{"type":70,"value":5686},"57 M+ grasps",{"type":70,"value":5688}," over 8515 Objaverse-XL (LVIS) objects",{"type":64,"tag":1964,"props":5690,"children":5691},{},[5692,5697],{"type":64,"tag":1986,"props":5693,"children":5694},{},[5695],{"type":70,"value":5696},"Grippers",{"type":64,"tag":1986,"props":5698,"children":5699},{},[5700],{"type":70,"value":5701},"Franka Panda, Robotiq-2f-140, suction (30 mm radius)",{"type":64,"tag":1964,"props":5703,"children":5704},{},[5705,5709],{"type":64,"tag":1986,"props":5706,"children":5707},{},[5708],{"type":70,"value":2615},{"type":64,"tag":1986,"props":5710,"children":5711},{},[5712,5714,5720],{"type":70,"value":5713},"WebDataset shards (",{"type":64,"tag":86,"props":5715,"children":5717},{"className":5716},[],[5718],{"type":70,"value":5719},"grasp_data\u002F{franka,robotiq2f140,suction}\u002Fshard_{0-7}.tar",{"type":70,"value":5721},") + train\u002Fvalid splits",{"type":64,"tag":1964,"props":5723,"children":5724},{},[5725,5729],{"type":64,"tag":1986,"props":5726,"children":5727},{},[5728],{"type":70,"value":2628},{"type":64,"tag":1986,"props":5730,"children":5731},{},[5732],{"type":70,"value":3712},{"type":64,"tag":1964,"props":5734,"children":5735},{},[5736,5740],{"type":64,"tag":1986,"props":5737,"children":5738},{},[5739],{"type":70,"value":4339},{"type":64,"tag":1986,"props":5741,"children":5742},{},[5743],{"type":64,"tag":206,"props":5744,"children":5747},{"href":5745,"rel":5746},"https:\u002F\u002Fhuggingface.co\u002Fadithyamurali\u002FGraspGenModels",[209],[5748],{"type":64,"tag":86,"props":5749,"children":5751},{"className":5750},[],[5752],{"type":70,"value":5753},"adithyamurali\u002FGraspGenModels",{"type":64,"tag":1964,"props":5755,"children":5756},{},[5757,5762],{"type":64,"tag":1986,"props":5758,"children":5759},{},[5760],{"type":70,"value":5761},"Tip",{"type":64,"tag":1986,"props":5763,"children":5764},{},[5765,5767,5772,5774],{"type":70,"value":5766},"Objaverse meshes are ",{"type":64,"tag":129,"props":5768,"children":5769},{},[5770],{"type":70,"value":5771},"NOT",{"type":70,"value":5773}," included — pull separately via the bundled ",{"type":64,"tag":86,"props":5775,"children":5777},{"className":5776},[],[5778],{"type":70,"value":5779},"download_objaverse.py",{"type":64,"tag":73,"props":5781,"children":5783},{"id":5782},"robotics-physical-material-properties",[5784],{"type":70,"value":765},{"type":64,"tag":80,"props":5786,"children":5787},{},[5788],{"type":70,"value":5789},"VoMP = Volumetric Mechanical Properties.",{"type":64,"tag":319,"props":5791,"children":5793},{"id":5792},"physicalai-robotics-physicalassets-vomp",[5794],{"type":70,"value":2498},{"type":64,"tag":1956,"props":5796,"children":5797},{},[5798,5812],{"type":64,"tag":1960,"props":5799,"children":5800},{},[5801],{"type":64,"tag":1964,"props":5802,"children":5803},{},[5804,5808],{"type":64,"tag":1968,"props":5805,"children":5806},{},[5807],{"type":70,"value":2533},{"type":64,"tag":1968,"props":5809,"children":5810},{},[5811],{"type":70,"value":2538},{"type":64,"tag":1979,"props":5813,"children":5814},{},[5815,5831,5843,5855],{"type":64,"tag":1964,"props":5816,"children":5817},{},[5818,5822],{"type":64,"tag":1986,"props":5819,"children":5820},{},[5821],{"type":70,"value":2549},{"type":64,"tag":1986,"props":5823,"children":5824},{},[5825],{"type":64,"tag":206,"props":5826,"children":5829},{"href":5827,"rel":5828},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-PhysicalAssets-VoMP",[209],[5830],{"type":70,"value":5827},{"type":64,"tag":1964,"props":5832,"children":5833},{},[5834,5838],{"type":64,"tag":1986,"props":5835,"children":5836},{},[5837],{"type":70,"value":2566},{"type":64,"tag":1986,"props":5839,"children":5840},{},[5841],{"type":70,"value":5842},"65.9 GB on-disk (full data nominally 125 GB pre-compression)",{"type":64,"tag":1964,"props":5844,"children":5845},{},[5846,5850],{"type":64,"tag":1986,"props":5847,"children":5848},{},[5849],{"type":70,"value":4762},{"type":64,"tag":1986,"props":5851,"children":5852},{},[5853],{"type":70,"value":5854},"1664 objects, 37,337,952 voxels, multi-view renders + VLM material annotations",{"type":64,"tag":1964,"props":5856,"children":5857},{},[5858,5862],{"type":64,"tag":1986,"props":5859,"children":5860},{},[5861],{"type":70,"value":2628},{"type":64,"tag":1986,"props":5863,"children":5864},{},[5865],{"type":70,"value":3176},{"type":64,"tag":319,"props":5867,"children":5869},{"id":5868},"physicalai-robotics-physicalassets-vomp-eval",[5870],{"type":70,"value":5871},"PhysicalAI-Robotics-PhysicalAssets-VoMP-Eval",{"type":64,"tag":1956,"props":5873,"children":5874},{},[5875,5889],{"type":64,"tag":1960,"props":5876,"children":5877},{},[5878],{"type":64,"tag":1964,"props":5879,"children":5880},{},[5881,5885],{"type":64,"tag":1968,"props":5882,"children":5883},{},[5884],{"type":70,"value":2533},{"type":64,"tag":1968,"props":5886,"children":5887},{},[5888],{"type":70,"value":2538},{"type":64,"tag":1979,"props":5890,"children":5891},{},[5892,5908,5920,5932],{"type":64,"tag":1964,"props":5893,"children":5894},{},[5895,5899],{"type":64,"tag":1986,"props":5896,"children":5897},{},[5898],{"type":70,"value":2549},{"type":64,"tag":1986,"props":5900,"children":5901},{},[5902],{"type":64,"tag":206,"props":5903,"children":5906},{"href":5904,"rel":5905},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Robotics-PhysicalAssets-VoMP-Eval",[209],[5907],{"type":70,"value":5904},{"type":64,"tag":1964,"props":5909,"children":5910},{},[5911,5915],{"type":64,"tag":1986,"props":5912,"children":5913},{},[5914],{"type":70,"value":2566},{"type":64,"tag":1986,"props":5916,"children":5917},{},[5918],{"type":70,"value":5919},"8.41 GB on-disk (full eval data nominally 125 GB pre-compression)",{"type":64,"tag":1964,"props":5921,"children":5922},{},[5923,5927],{"type":64,"tag":1986,"props":5924,"children":5925},{},[5926],{"type":70,"value":4775},{"type":64,"tag":1986,"props":5928,"children":5929},{},[5930],{"type":70,"value":5931},"Held-out eval split for the VoMP model",{"type":64,"tag":1964,"props":5933,"children":5934},{},[5935,5939],{"type":64,"tag":1986,"props":5936,"children":5937},{},[5938],{"type":70,"value":2628},{"type":64,"tag":1986,"props":5940,"children":5941},{},[5942],{"type":70,"value":3176},{"type":64,"tag":73,"props":5944,"children":5946},{"id":5945},"spatial-intelligence-simready-scenes",[5947],{"type":70,"value":775},{"type":64,"tag":319,"props":5949,"children":5951},{"id":5950},"physicalai-simready-warehouse-01",[5952],{"type":70,"value":2330},{"type":64,"tag":80,"props":5954,"children":5955},{},[5956],{"type":70,"value":5957},"OpenUSD warehouse scene + asset library.",{"type":64,"tag":1956,"props":5959,"children":5960},{},[5961,5975],{"type":64,"tag":1960,"props":5962,"children":5963},{},[5964],{"type":64,"tag":1964,"props":5965,"children":5966},{},[5967,5971],{"type":64,"tag":1968,"props":5968,"children":5969},{},[5970],{"type":70,"value":2533},{"type":64,"tag":1968,"props":5972,"children":5973},{},[5974],{"type":70,"value":2538},{"type":64,"tag":1979,"props":5976,"children":5977},{},[5978,5994,6013,6026,6039,6057,6068],{"type":64,"tag":1964,"props":5979,"children":5980},{},[5981,5985],{"type":64,"tag":1986,"props":5982,"children":5983},{},[5984],{"type":70,"value":2549},{"type":64,"tag":1986,"props":5986,"children":5987},{},[5988],{"type":64,"tag":206,"props":5989,"children":5992},{"href":5990,"rel":5991},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-SimReady-Warehouse-01",[209],[5993],{"type":70,"value":5990},{"type":64,"tag":1964,"props":5995,"children":5996},{},[5997,6001],{"type":64,"tag":1986,"props":5998,"children":5999},{},[6000],{"type":70,"value":2566},{"type":64,"tag":1986,"props":6002,"children":6003},{},[6004,6006,6012],{"type":70,"value":6005},"14.4 GB, 753 USD assets + master scene (",{"type":64,"tag":86,"props":6007,"children":6009},{"className":6008},[],[6010],{"type":70,"value":6011},"physical_ai_simready_warehouse_01.usd",{"type":70,"value":2019},{"type":64,"tag":1964,"props":6014,"children":6015},{},[6016,6021],{"type":64,"tag":1986,"props":6017,"children":6018},{},[6019],{"type":70,"value":6020},"Asset class",{"type":64,"tag":1986,"props":6022,"children":6023},{},[6024],{"type":70,"value":6025},"Prop \u002F Assembly \u002F Scenario; 1.1.0 adds physically-graspable subset",{"type":64,"tag":1964,"props":6027,"children":6028},{},[6029,6034],{"type":64,"tag":1986,"props":6030,"children":6031},{},[6032],{"type":70,"value":6033},"Metadata",{"type":64,"tag":1986,"props":6035,"children":6036},{},[6037],{"type":70,"value":6038},"CSV catalogue with WikiData Q-codes, mass (kg), thumbnails",{"type":64,"tag":1964,"props":6040,"children":6041},{},[6042,6047],{"type":64,"tag":1986,"props":6043,"children":6044},{},[6045],{"type":70,"value":6046},"Target",{"type":64,"tag":1986,"props":6048,"children":6049},{},[6050,6055],{"type":64,"tag":129,"props":6051,"children":6052},{},[6053],{"type":70,"value":6054},"Isaac Sim 4.x",{"type":70,"value":6056}," (Properties → disable Instanceable → Physics → Rigid Body to make assets dynamic)",{"type":64,"tag":1964,"props":6058,"children":6059},{},[6060,6064],{"type":64,"tag":1986,"props":6061,"children":6062},{},[6063],{"type":70,"value":2628},{"type":64,"tag":1986,"props":6065,"children":6066},{},[6067],{"type":70,"value":3176},{"type":64,"tag":1964,"props":6069,"children":6070},{},[6071,6075],{"type":64,"tag":1986,"props":6072,"children":6073},{},[6074],{"type":70,"value":2717},{"type":64,"tag":1986,"props":6076,"children":6077},{},[6078],{"type":70,"value":6079},"Upstream consumers without an in-repo skill: Isaac Sim \u002F Isaac Lab, Omniverse SDG.",{"type":64,"tag":319,"props":6081,"children":6083},{"id":6082},"physicalai-digitalcousin-assets",[6084],{"type":70,"value":2349},{"type":64,"tag":80,"props":6086,"children":6087},{},[6088],{"type":70,"value":6089},"Companion-asset library for the GR1 tabletop sim environments.",{"type":64,"tag":1956,"props":6091,"children":6092},{},[6093,6107],{"type":64,"tag":1960,"props":6094,"children":6095},{},[6096],{"type":64,"tag":1964,"props":6097,"children":6098},{},[6099,6103],{"type":64,"tag":1968,"props":6100,"children":6101},{},[6102],{"type":70,"value":2533},{"type":64,"tag":1968,"props":6104,"children":6105},{},[6106],{"type":70,"value":2538},{"type":64,"tag":1979,"props":6108,"children":6109},{},[6110,6126,6138,6151],{"type":64,"tag":1964,"props":6111,"children":6112},{},[6113,6117],{"type":64,"tag":1986,"props":6114,"children":6115},{},[6116],{"type":70,"value":2549},{"type":64,"tag":1986,"props":6118,"children":6119},{},[6120],{"type":64,"tag":206,"props":6121,"children":6124},{"href":6122,"rel":6123},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-DigitalCousin-Assets",[209],[6125],{"type":70,"value":6122},{"type":64,"tag":1964,"props":6127,"children":6128},{},[6129,6133],{"type":64,"tag":1986,"props":6130,"children":6131},{},[6132],{"type":70,"value":2566},{"type":64,"tag":1986,"props":6134,"children":6135},{},[6136],{"type":70,"value":6137},"270 MB",{"type":64,"tag":1964,"props":6139,"children":6140},{},[6141,6146],{"type":64,"tag":1986,"props":6142,"children":6143},{},[6144],{"type":70,"value":6145},"Content",{"type":64,"tag":1986,"props":6147,"children":6148},{},[6149],{"type":70,"value":6150},"3D meshes, textures, metadata for tabletop objects (mug, bottle, bowl, container, …) used by GR1 sim tasks",{"type":64,"tag":1964,"props":6152,"children":6153},{},[6154,6158],{"type":64,"tag":1986,"props":6155,"children":6156},{},[6157],{"type":70,"value":2628},{"type":64,"tag":1986,"props":6159,"children":6160},{},[6161,6165],{"type":64,"tag":129,"props":6162,"children":6163},{},[6164],{"type":70,"value":4570},{"type":70,"value":6166}," (non-commercial)",{"type":64,"tag":319,"props":6168,"children":6170},{"id":6169},"physicalai-smartspaces",[6171],{"type":70,"value":2385},{"type":64,"tag":80,"props":6173,"children":6174},{},[6175],{"type":70,"value":6176},"Multi-camera tracking + 3D box benchmark (AI City Challenge).",{"type":64,"tag":1956,"props":6178,"children":6179},{},[6180,6194],{"type":64,"tag":1960,"props":6181,"children":6182},{},[6183],{"type":64,"tag":1964,"props":6184,"children":6185},{},[6186,6190],{"type":64,"tag":1968,"props":6187,"children":6188},{},[6189],{"type":70,"value":2533},{"type":64,"tag":1968,"props":6191,"children":6192},{},[6193],{"type":70,"value":2538},{"type":64,"tag":1979,"props":6195,"children":6196},{},[6197,6213,6225,6238,6251,6264,6290],{"type":64,"tag":1964,"props":6198,"children":6199},{},[6200,6204],{"type":64,"tag":1986,"props":6201,"children":6202},{},[6203],{"type":70,"value":2549},{"type":64,"tag":1986,"props":6205,"children":6206},{},[6207],{"type":64,"tag":206,"props":6208,"children":6211},{"href":6209,"rel":6210},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-SmartSpaces",[209],[6212],{"type":70,"value":6209},{"type":64,"tag":1964,"props":6214,"children":6215},{},[6216,6220],{"type":64,"tag":1986,"props":6217,"children":6218},{},[6219],{"type":70,"value":2566},{"type":64,"tag":1986,"props":6221,"children":6222},{},[6223],{"type":70,"value":6224},"3.53 TB total (216 GB MTMC_Tracking_2024 + 3.31 TB MTMC_Tracking_2025)",{"type":64,"tag":1964,"props":6226,"children":6227},{},[6228,6233],{"type":64,"tag":1986,"props":6229,"children":6230},{},[6231],{"type":70,"value":6232},"2024",{"type":64,"tag":1986,"props":6234,"children":6235},{},[6236],{"type":70,"value":6237},"90 scenes, 212 h, 953 cameras — Person-only, 2D boxes + multi-cam IDs (52M \u002F 135M)",{"type":64,"tag":1964,"props":6239,"children":6240},{},[6241,6246],{"type":64,"tag":1986,"props":6242,"children":6243},{},[6244],{"type":70,"value":6245},"2025",{"type":64,"tag":1986,"props":6247,"children":6248},{},[6249],{"type":70,"value":6250},"23 scenes, 42 h, 504 cameras — Person\u002FForklift\u002FNovaCarter\u002FTransporter\u002FFourierGR1T2\u002FAgilityDigit, 3D boxes + depth maps (8.9M \u002F 73M)",{"type":64,"tag":1964,"props":6252,"children":6253},{},[6254,6259],{"type":64,"tag":1986,"props":6255,"children":6256},{},[6257],{"type":70,"value":6258},"Splits",{"type":64,"tag":1986,"props":6260,"children":6261},{},[6262],{"type":70,"value":6263},"Warehouse (train\u002Fval\u002Ftest) + Lab (val) + Hospital (val) + 4 test scenes added 2025-05-28",{"type":64,"tag":1964,"props":6265,"children":6266},{},[6267,6272],{"type":64,"tag":1986,"props":6268,"children":6269},{},[6270],{"type":70,"value":6271},"Eval",{"type":64,"tag":1986,"props":6273,"children":6274},{},[6275,6281,6282,6288],{"type":64,"tag":206,"props":6276,"children":6279},{"href":6277,"rel":6278},"https:\u002F\u002Feval.aicitychallenge.org\u002Faicity2024",[209],[6280],{"type":70,"value":6277},{"type":70,"value":4674},{"type":64,"tag":206,"props":6283,"children":6286},{"href":6284,"rel":6285},"https:\u002F\u002Feval.aicitychallenge.org\u002Faicity2025",[209],[6287],{"type":70,"value":6284},{"type":70,"value":6289},"; 3D-bbox HOTA metric for 2025",{"type":64,"tag":1964,"props":6291,"children":6292},{},[6293,6297],{"type":64,"tag":1986,"props":6294,"children":6295},{},[6296],{"type":70,"value":3203},{"type":64,"tag":1986,"props":6298,"children":6299},{},[6300,6306],{"type":64,"tag":206,"props":6301,"children":6304},{"href":6302,"rel":6303},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2412.00692",[209],[6305],{"type":70,"value":6302},{"type":70,"value":6307}," (MCBLT)",{"type":64,"tag":319,"props":6309,"children":6311},{"id":6310},"physicalai-spatial-intelligence-warehouse",[6312],{"type":70,"value":2404},{"type":64,"tag":80,"props":6314,"children":6315},{},[6316],{"type":70,"value":6317},"VLM-style spatial QA in warehouses.",{"type":64,"tag":1956,"props":6319,"children":6320},{},[6321,6335],{"type":64,"tag":1960,"props":6322,"children":6323},{},[6324],{"type":64,"tag":1964,"props":6325,"children":6326},{},[6327,6331],{"type":64,"tag":1968,"props":6328,"children":6329},{},[6330],{"type":70,"value":2533},{"type":64,"tag":1968,"props":6332,"children":6333},{},[6334],{"type":70,"value":2538},{"type":64,"tag":1979,"props":6336,"children":6337},{},[6338,6354,6366,6408,6421,6434,6445,6456],{"type":64,"tag":1964,"props":6339,"children":6340},{},[6341,6345],{"type":64,"tag":1986,"props":6342,"children":6343},{},[6344],{"type":70,"value":2549},{"type":64,"tag":1986,"props":6346,"children":6347},{},[6348],{"type":64,"tag":206,"props":6349,"children":6352},{"href":6350,"rel":6351},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-Spatial-Intelligence-Warehouse",[209],[6353],{"type":70,"value":6350},{"type":64,"tag":1964,"props":6355,"children":6356},{},[6357,6361],{"type":64,"tag":1986,"props":6358,"children":6359},{},[6360],{"type":70,"value":2566},{"type":64,"tag":1986,"props":6362,"children":6363},{},[6364],{"type":70,"value":6365},"261 GB",{"type":64,"tag":1964,"props":6367,"children":6368},{},[6369,6374],{"type":64,"tag":1986,"props":6370,"children":6371},{},[6372],{"type":70,"value":6373},"QA pairs",{"type":64,"tag":1986,"props":6375,"children":6376},{},[6377,6379,6385,6386,6392,6393,6399,6401,6407],{"type":70,"value":6378},"499k train + 19k test + 1.9k val (categories: ",{"type":64,"tag":86,"props":6380,"children":6382},{"className":6381},[],[6383],{"type":70,"value":6384},"left_right",{"type":70,"value":101},{"type":64,"tag":86,"props":6387,"children":6389},{"className":6388},[],[6390],{"type":70,"value":6391},"multi_choice_question",{"type":70,"value":101},{"type":64,"tag":86,"props":6394,"children":6396},{"className":6395},[],[6397],{"type":70,"value":6398},"distance",{"type":70,"value":6400}," in metres, ",{"type":64,"tag":86,"props":6402,"children":6404},{"className":6403},[],[6405],{"type":70,"value":6406},"count",{"type":70,"value":2019},{"type":64,"tag":1964,"props":6409,"children":6410},{},[6411,6416],{"type":64,"tag":1986,"props":6412,"children":6413},{},[6414],{"type":70,"value":6415},"Imagery",{"type":64,"tag":1986,"props":6417,"children":6418},{},[6419],{"type":70,"value":6420},"~95k RGB-D pairs, RLE object masks (pycoco), LLaVA-style conversations",{"type":64,"tag":1964,"props":6422,"children":6423},{},[6424,6429],{"type":64,"tag":1986,"props":6425,"children":6426},{},[6427],{"type":70,"value":6428},"Annotation",{"type":64,"tag":1986,"props":6430,"children":6431},{},[6432],{"type":70,"value":6433},"Rule-based + Llama-3.1-70B-Instruct refinement",{"type":64,"tag":1964,"props":6435,"children":6436},{},[6437,6441],{"type":64,"tag":1986,"props":6438,"children":6439},{},[6440],{"type":70,"value":5294},{"type":64,"tag":1986,"props":6442,"children":6443},{},[6444],{"type":70,"value":5299},{"type":64,"tag":1964,"props":6446,"children":6447},{},[6448,6452],{"type":64,"tag":1986,"props":6449,"children":6450},{},[6451],{"type":70,"value":2628},{"type":64,"tag":1986,"props":6453,"children":6454},{},[6455],{"type":70,"value":3176},{"type":64,"tag":1964,"props":6457,"children":6458},{},[6459,6463],{"type":64,"tag":1986,"props":6460,"children":6461},{},[6462],{"type":70,"value":2615},{"type":64,"tag":1986,"props":6464,"children":6465},{},[6466,6472,6474,6480,6481,6487],{"type":64,"tag":86,"props":6467,"children":6469},{"className":6468},[],[6470],{"type":70,"value":6471},"train.json",{"type":70,"value":6473}," \u002F ",{"type":64,"tag":86,"props":6475,"children":6477},{"className":6476},[],[6478],{"type":70,"value":6479},"val.json",{"type":70,"value":6473},{"type":64,"tag":86,"props":6482,"children":6484},{"className":6483},[],[6485],{"type":70,"value":6486},"test.json",{"type":70,"value":6488}," + chunked TAR-GZs of images + depths",{"type":64,"tag":319,"props":6490,"children":6492},{"id":6491},"physicalai-spatialintelligence-lyra-sdg",[6493],{"type":70,"value":1546},{"type":64,"tag":80,"props":6495,"children":6496},{},[6497,6499,6505],{"type":70,"value":6498},"GEN3C-derived multi-view 3D + 4D training data for ",{"type":64,"tag":86,"props":6500,"children":6502},{"className":6501},[],[6503],{"type":70,"value":6504},"nv-tlabs\u002Flyra",{"type":70,"value":621},{"type":64,"tag":1956,"props":6507,"children":6508},{},[6509,6523],{"type":64,"tag":1960,"props":6510,"children":6511},{},[6512],{"type":64,"tag":1964,"props":6513,"children":6514},{},[6515,6519],{"type":64,"tag":1968,"props":6516,"children":6517},{},[6518],{"type":70,"value":2533},{"type":64,"tag":1968,"props":6520,"children":6521},{},[6522],{"type":70,"value":2538},{"type":64,"tag":1979,"props":6524,"children":6525},{},[6526,6542,6557,6569,6589,6600,6618],{"type":64,"tag":1964,"props":6527,"children":6528},{},[6529,6533],{"type":64,"tag":1986,"props":6530,"children":6531},{},[6532],{"type":70,"value":2549},{"type":64,"tag":1986,"props":6534,"children":6535},{},[6536],{"type":64,"tag":206,"props":6537,"children":6540},{"href":6538,"rel":6539},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FPhysicalAI-SpatialIntelligence-Lyra-SDG",[209],[6541],{"type":70,"value":6538},{"type":64,"tag":1964,"props":6543,"children":6544},{},[6545,6549],{"type":64,"tag":1986,"props":6546,"children":6547},{},[6548],{"type":70,"value":2566},{"type":64,"tag":1986,"props":6550,"children":6551},{},[6552],{"type":64,"tag":129,"props":6553,"children":6554},{},[6555],{"type":70,"value":6556},"25 TB",{"type":64,"tag":1964,"props":6558,"children":6559},{},[6560,6564],{"type":64,"tag":1986,"props":6561,"children":6562},{},[6563],{"type":70,"value":4160},{"type":64,"tag":1986,"props":6565,"children":6566},{},[6567],{"type":70,"value":6568},"59,031 multi-view 3D examples (354,186 videos) + 7,378 4D examples (44,268 videos), 6 trajectories per source",{"type":64,"tag":1964,"props":6570,"children":6571},{},[6572,6576],{"type":64,"tag":1986,"props":6573,"children":6574},{},[6575],{"type":70,"value":3130},{"type":64,"tag":1986,"props":6577,"children":6578},{},[6579,6581,6587],{"type":70,"value":6580},"RGB MP4 + camera pose ",{"type":64,"tag":86,"props":6582,"children":6584},{"className":6583},[],[6585],{"type":70,"value":6586},".npz",{"type":70,"value":6588}," + depth zip",{"type":64,"tag":1964,"props":6590,"children":6591},{},[6592,6596],{"type":64,"tag":1986,"props":6593,"children":6594},{},[6595],{"type":70,"value":2628},{"type":64,"tag":1986,"props":6597,"children":6598},{},[6599],{"type":70,"value":3176},{"type":64,"tag":1964,"props":6601,"children":6602},{},[6603,6607],{"type":64,"tag":1986,"props":6604,"children":6605},{},[6606],{"type":70,"value":3203},{"type":64,"tag":1986,"props":6608,"children":6609},{},[6610,6616],{"type":64,"tag":206,"props":6611,"children":6614},{"href":6612,"rel":6613},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2509.19296",[209],[6615],{"type":70,"value":6612},{"type":70,"value":6617}," (Lyra)",{"type":64,"tag":1964,"props":6619,"children":6620},{},[6621,6626],{"type":64,"tag":1986,"props":6622,"children":6623},{},[6624],{"type":70,"value":6625},"Pair with",{"type":64,"tag":1986,"props":6627,"children":6628},{},[6629,6640,6642,6652],{"type":64,"tag":206,"props":6630,"children":6633},{"href":6631,"rel":6632},"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002Fnvidia\u002FLyra-Testing-Example",[209],[6634],{"type":64,"tag":86,"props":6635,"children":6637},{"className":6636},[],[6638],{"type":70,"value":6639},"nvidia\u002FLyra-Testing-Example",{"type":70,"value":6641}," for inference; ",{"type":64,"tag":206,"props":6643,"children":6646},{"href":6644,"rel":6645},"https:\u002F\u002Fgithub.com\u002Fnv-tlabs\u002Flyra",[209],[6647],{"type":64,"tag":86,"props":6648,"children":6650},{"className":6649},[],[6651],{"type":70,"value":6504},{"type":70,"value":6653}," for training",{"type":64,"tag":319,"props":6655,"children":6657},{"id":6656},"see-also-physicalai-nurec-ppisp",[6658],{"type":70,"value":6659},"(See also) PhysicalAI-NuRec-PPISP",{"type":64,"tag":80,"props":6661,"children":6662},{},[6663,6665,6670],{"type":70,"value":6664},"Listed in ",{"type":64,"tag":206,"props":6666,"children":6667},{"href":730},[6668],{"type":70,"value":6669},"§ 7",{"type":70,"value":6671}," but is also a\nspatial-intelligence radiance-field benchmark.",{"type":64,"tag":73,"props":6673,"children":6675},{"id":6674},"license-decision-tree",[6676],{"type":70,"value":795},{"type":64,"tag":1956,"props":6678,"children":6679},{},[6680,6704],{"type":64,"tag":1960,"props":6681,"children":6682},{},[6683],{"type":64,"tag":1964,"props":6684,"children":6685},{},[6686,6690,6695,6700],{"type":64,"tag":1968,"props":6687,"children":6688},{},[6689],{"type":70,"value":2628},{"type":64,"tag":1968,"props":6691,"children":6692},{},[6693],{"type":70,"value":6694},"Commercial OK?",{"type":64,"tag":1968,"props":6696,"children":6697},{},[6698],{"type":70,"value":6699},"Reproducible?",{"type":64,"tag":1968,"props":6701,"children":6702},{},[6703],{"type":70,"value":17},{"type":64,"tag":1979,"props":6705,"children":6706},{},[6707,6732,6765],{"type":64,"tag":1964,"props":6708,"children":6709},{},[6710,6717,6722,6727],{"type":64,"tag":1986,"props":6711,"children":6712},{},[6713],{"type":64,"tag":129,"props":6714,"children":6715},{},[6716],{"type":70,"value":3176},{"type":64,"tag":1986,"props":6718,"children":6719},{},[6720],{"type":70,"value":6721},"✅ (with attribution)",{"type":64,"tag":1986,"props":6723,"children":6724},{},[6725],{"type":70,"value":6726},"✅ (must keep notice)",{"type":64,"tag":1986,"props":6728,"children":6729},{},[6730],{"type":70,"value":6731},"most — Cosmos-Drive-Dreams, GraspGen, all Manipulation-* (except where noted), GR00T-Teleop-G1, GR00T-Tuned-Tasks, GR00T-GR1, GR00T-Eval, GR00T-X-Embodiment-Sim, Robotics-NuRec, NuRec-PPISP, Open-H-Embodiment, SmartSpaces (CC-BY-4.0 implied via card), Spatial-Intelligence-Warehouse, SpatialIntelligence-Lyra-SDG, SimReady-Warehouse-01, VoMP \u002F VoMP-Eval, Manipulation-Augmented",{"type":64,"tag":1964,"props":6733,"children":6734},{},[6735,6743,6748,6753],{"type":64,"tag":1986,"props":6736,"children":6737},{},[6738,6742],{"type":64,"tag":129,"props":6739,"children":6740},{},[6741],{"type":70,"value":4570},{"type":70,"value":6166},{"type":64,"tag":1986,"props":6744,"children":6745},{},[6746],{"type":70,"value":6747},"❌",{"type":64,"tag":1986,"props":6749,"children":6750},{},[6751],{"type":70,"value":6752},"✅ research",{"type":64,"tag":1986,"props":6754,"children":6755},{},[6756,6758,6763],{"type":70,"value":6757},"GR00T-Teleop-Sim, DigitalCousin-Assets, ",{"type":64,"tag":129,"props":6759,"children":6760},{},[6761],{"type":70,"value":6762},"all 4 mindmap",{"type":70,"value":6764}," datasets",{"type":64,"tag":1964,"props":6766,"children":6767},{},[6768,6777,6782,6787],{"type":64,"tag":1986,"props":6769,"children":6770},{},[6771,6775],{"type":64,"tag":129,"props":6772,"children":6773},{},[6774],{"type":70,"value":2636},{"type":70,"value":6776}," (gated, AV-only purpose, 12-month expiry)",{"type":64,"tag":1986,"props":6778,"children":6779},{},[6780],{"type":70,"value":6781},"✅ ONLY for AV \u002F ADAS development on NVIDIA tech",{"type":64,"tag":1986,"props":6783,"children":6784},{},[6785],{"type":70,"value":6786},"❌ — no derivative works, no redistribution, no biometric \u002F re-id \u002F surveillance use",{"type":64,"tag":1986,"props":6788,"children":6789},{},[6790,6795,6796,6802,6803],{"type":64,"tag":86,"props":6791,"children":6793},{"className":6792},[],[6794],{"type":70,"value":667},{"type":70,"value":101},{"type":64,"tag":86,"props":6797,"children":6799},{"className":6798},[],[6800],{"type":70,"value":6801},"…-NCore",{"type":70,"value":101},{"type":64,"tag":86,"props":6804,"children":6806},{"className":6805},[],[6807],{"type":70,"value":6808},"…-NuRec",{"type":64,"tag":80,"props":6810,"children":6811},{},[6812],{"type":70,"value":6813},"For internal NVIDIA use, the auto-derivable rule of thumb:",{"type":64,"tag":629,"props":6815,"children":6816},{},[6817,6836,6848,6860],{"type":64,"tag":156,"props":6818,"children":6819},{},[6820,6822,6828,6829,6834],{"type":70,"value":6821},"If ",{"type":64,"tag":86,"props":6823,"children":6825},{"className":6824},[],[6826],{"type":70,"value":6827},"Robotics-Manipulation-*",{"type":70,"value":1396},{"type":64,"tag":129,"props":6830,"children":6831},{},[6832],{"type":70,"value":6833},"not",{"type":70,"value":6835}," mindmap \u002F DigitalCousin-Assets → CC-BY-4.0 commercial OK.",{"type":64,"tag":156,"props":6837,"children":6838},{},[6839,6840,6846],{"type":70,"value":6821},{"type":64,"tag":86,"props":6841,"children":6843},{"className":6842},[],[6844],{"type":70,"value":6845},"mindmap-*",{"type":70,"value":6847}," → research only (NC).",{"type":64,"tag":156,"props":6849,"children":6850},{},[6851,6852,6858],{"type":70,"value":6821},{"type":64,"tag":86,"props":6853,"children":6855},{"className":6854},[],[6856],{"type":70,"value":6857},"Autonomous-Vehicles*",{"type":70,"value":6859}," → AV License only, gated.",{"type":64,"tag":156,"props":6861,"children":6862},{},[6863],{"type":70,"value":6864},"Everything else → check the card.",{"type":64,"tag":73,"props":6866,"children":6868},{"id":6867},"cross-skill-usage-map",[6869],{"type":70,"value":806},{"type":64,"tag":80,"props":6871,"children":6872},{},[6873],{"type":70,"value":6874},"Sibling skills in this hub are linked by relative path; upstream\nprojects without an in-repo skill are linked by URL.",{"type":64,"tag":1956,"props":6876,"children":6877},{},[6878,6898],{"type":64,"tag":1960,"props":6879,"children":6880},{},[6881],{"type":64,"tag":1964,"props":6882,"children":6883},{},[6884,6888,6893],{"type":64,"tag":1968,"props":6885,"children":6886},{},[6887],{"type":70,"value":4931},{"type":64,"tag":1968,"props":6889,"children":6890},{},[6891],{"type":70,"value":6892},"In-repo sibling skill(s)",{"type":64,"tag":1968,"props":6894,"children":6895},{},[6896],{"type":70,"value":6897},"Upstream consumers (no in-repo skill)",{"type":64,"tag":1979,"props":6899,"children":6900},{},[6901,6939,6975,7011,7048,7069,7104,7130,7158,7197,7241,7270,7300,7337,7366,7396,7417,7444,7465,7486,7507,7527],{"type":64,"tag":1964,"props":6902,"children":6903},{},[6904,6912,6923],{"type":64,"tag":1986,"props":6905,"children":6906},{},[6907],{"type":64,"tag":86,"props":6908,"children":6910},{"className":6909},[],[6911],{"type":70,"value":667},{"type":64,"tag":1986,"props":6913,"children":6914},{},[6915],{"type":64,"tag":206,"props":6916,"children":6917},{"href":2011},[6918],{"type":64,"tag":86,"props":6919,"children":6921},{"className":6920},[],[6922],{"type":70,"value":2011},{"type":64,"tag":1986,"props":6924,"children":6925},{},[6926,6931,6932,6937],{"type":64,"tag":86,"props":6927,"children":6929},{"className":6928},[],[6930],{"type":70,"value":2755},{"type":70,"value":101},{"type":64,"tag":86,"props":6933,"children":6935},{"className":6934},[],[6936],{"type":70,"value":2762},{"type":70,"value":6938},", CARLA",{"type":64,"tag":1964,"props":6940,"children":6941},{},[6942,6950,6970],{"type":64,"tag":1986,"props":6943,"children":6944},{},[6945],{"type":64,"tag":86,"props":6946,"children":6948},{"className":6947},[],[6949],{"type":70,"value":6801},{"type":64,"tag":1986,"props":6951,"children":6952},{},[6953,6961,6962],{"type":64,"tag":206,"props":6954,"children":6955},{"href":2011},[6956],{"type":64,"tag":86,"props":6957,"children":6959},{"className":6958},[],[6960],{"type":70,"value":2011},{"type":70,"value":101},{"type":64,"tag":206,"props":6963,"children":6964},{"href":2733},[6965],{"type":64,"tag":86,"props":6966,"children":6968},{"className":6967},[],[6969],{"type":70,"value":2733},{"type":64,"tag":1986,"props":6971,"children":6972},{},[6973],{"type":70,"value":6974},"—",{"type":64,"tag":1964,"props":6976,"children":6977},{},[6978,6986,7006],{"type":64,"tag":1986,"props":6979,"children":6980},{},[6981],{"type":64,"tag":86,"props":6982,"children":6984},{"className":6983},[],[6985],{"type":70,"value":6808},{"type":64,"tag":1986,"props":6987,"children":6988},{},[6989,6997,6998],{"type":64,"tag":206,"props":6990,"children":6991},{"href":3037},[6992],{"type":64,"tag":86,"props":6993,"children":6995},{"className":6994},[],[6996],{"type":70,"value":3037},{"type":70,"value":101},{"type":64,"tag":206,"props":6999,"children":7000},{"href":3056},[7001],{"type":64,"tag":86,"props":7002,"children":7004},{"className":7003},[],[7005],{"type":70,"value":3056},{"type":64,"tag":1986,"props":7007,"children":7008},{},[7009],{"type":70,"value":7010},"CARLA (NuRec integration 0.9.16+)",{"type":64,"tag":1964,"props":7012,"children":7013},{},[7014,7023,7027],{"type":64,"tag":1986,"props":7015,"children":7016},{},[7017],{"type":64,"tag":86,"props":7018,"children":7020},{"className":7019},[],[7021],{"type":70,"value":7022},"…-Cosmos-Drive-Dreams",{"type":64,"tag":1986,"props":7024,"children":7025},{},[7026],{"type":70,"value":6974},{"type":64,"tag":1986,"props":7028,"children":7029},{},[7030,7035,7036,7041,7042,7047],{"type":64,"tag":86,"props":7031,"children":7033},{"className":7032},[],[7034],{"type":70,"value":3230},{"type":70,"value":101},{"type":64,"tag":86,"props":7037,"children":7039},{"className":7038},[],[7040],{"type":70,"value":3237},{"type":70,"value":101},{"type":64,"tag":86,"props":7043,"children":7045},{"className":7044},[],[7046],{"type":70,"value":2762},{"type":70,"value":6938},{"type":64,"tag":1964,"props":7049,"children":7050},{},[7051,7060,7064],{"type":64,"tag":1986,"props":7052,"children":7053},{},[7054],{"type":64,"tag":86,"props":7055,"children":7057},{"className":7056},[],[7058],{"type":70,"value":7059},"…-Cosmos-Synthetic",{"type":64,"tag":1986,"props":7061,"children":7062},{},[7063],{"type":70,"value":6974},{"type":64,"tag":1986,"props":7065,"children":7066},{},[7067],{"type":70,"value":7068},"(pointer to Cosmos-Drive-Dreams)",{"type":64,"tag":1964,"props":7070,"children":7071},{},[7072,7095,7099],{"type":64,"tag":1986,"props":7073,"children":7074},{},[7075,7081,7082,7088,7089],{"type":64,"tag":86,"props":7076,"children":7078},{"className":7077},[],[7079],{"type":70,"value":7080},"Robotics-Manipulation-Kitchen",{"type":70,"value":6473},{"type":64,"tag":86,"props":7083,"children":7085},{"className":7084},[],[7086],{"type":70,"value":7087},"-Objects",{"type":70,"value":6473},{"type":64,"tag":86,"props":7090,"children":7092},{"className":7091},[],[7093],{"type":70,"value":7094},"-SingleArm",{"type":64,"tag":1986,"props":7096,"children":7097},{},[7098],{"type":70,"value":6974},{"type":64,"tag":1986,"props":7100,"children":7101},{},[7102],{"type":70,"value":7103},"Isaac Sim \u002F Isaac Lab",{"type":64,"tag":1964,"props":7105,"children":7106},{},[7107,7116,7120],{"type":64,"tag":1986,"props":7108,"children":7109},{},[7110],{"type":64,"tag":86,"props":7111,"children":7113},{"className":7112},[],[7114],{"type":70,"value":7115},"Robotics-Manipulation-Augmented",{"type":64,"tag":1986,"props":7117,"children":7118},{},[7119],{"type":70,"value":6974},{"type":64,"tag":1986,"props":7121,"children":7122},{},[7123,7128],{"type":64,"tag":86,"props":7124,"children":7126},{"className":7125},[],[7127],{"type":70,"value":3230},{"type":70,"value":7129}," (Transfer1 path), Isaac Sim \u002F Isaac Lab",{"type":64,"tag":1964,"props":7131,"children":7132},{},[7133,7149,7153],{"type":64,"tag":1986,"props":7134,"children":7135},{},[7136,7142,7143],{"type":64,"tag":86,"props":7137,"children":7139},{"className":7138},[],[7140],{"type":70,"value":7141},"Robotics-Manipulation-Kitchen-Demos",{"type":70,"value":4674},{"type":64,"tag":86,"props":7144,"children":7146},{"className":7145},[],[7147],{"type":70,"value":7148},"-Kitchen-MJCF",{"type":64,"tag":1986,"props":7150,"children":7151},{},[7152],{"type":70,"value":6974},{"type":64,"tag":1986,"props":7154,"children":7155},{},[7156],{"type":70,"value":7157},"MuJoCo direct; Isaac Sim for MJCF→USD",{"type":64,"tag":1964,"props":7159,"children":7160},{},[7161,7177,7181],{"type":64,"tag":1986,"props":7162,"children":7163},{},[7164,7170,7171],{"type":64,"tag":86,"props":7165,"children":7167},{"className":7166},[],[7168],{"type":70,"value":7169},"Robotics-GR00T-X-Embodiment-Sim",{"type":70,"value":6473},{"type":64,"tag":86,"props":7172,"children":7174},{"className":7173},[],[7175],{"type":70,"value":7176},"-Tuned-Tasks",{"type":64,"tag":1986,"props":7178,"children":7179},{},[7180],{"type":70,"value":6974},{"type":64,"tag":1986,"props":7182,"children":7183},{},[7184,7195],{"type":64,"tag":206,"props":7185,"children":7188},{"href":7186,"rel":7187},"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FIsaac-GR00T",[209],[7189],{"type":64,"tag":86,"props":7190,"children":7192},{"className":7191},[],[7193],{"type":70,"value":7194},"NVIDIA\u002FIsaac-GR00T",{"type":70,"value":7196},", Isaac Sim",{"type":64,"tag":1964,"props":7198,"children":7199},{},[7200,7225,7229],{"type":64,"tag":1986,"props":7201,"children":7202},{},[7203,7209,7210,7216,7217,7223],{"type":64,"tag":86,"props":7204,"children":7206},{"className":7205},[],[7207],{"type":70,"value":7208},"Robotics-GR00T-Teleop-G1",{"type":70,"value":6473},{"type":64,"tag":86,"props":7211,"children":7213},{"className":7212},[],[7214],{"type":70,"value":7215},"-Sim",{"type":70,"value":6473},{"type":64,"tag":86,"props":7218,"children":7220},{"className":7219},[],[7221],{"type":70,"value":7222},"-GR1",{"type":70,"value":7224}," (DreamDojo)",{"type":64,"tag":1986,"props":7226,"children":7227},{},[7228],{"type":70,"value":6974},{"type":64,"tag":1986,"props":7230,"children":7231},{},[7232],{"type":64,"tag":206,"props":7233,"children":7235},{"href":7186,"rel":7234},[209],[7236],{"type":64,"tag":86,"props":7237,"children":7239},{"className":7238},[],[7240],{"type":70,"value":7194},{"type":64,"tag":1964,"props":7242,"children":7243},{},[7244,7261,7265],{"type":64,"tag":1986,"props":7245,"children":7246},{},[7247,7253,7255],{"type":64,"tag":86,"props":7248,"children":7250},{"className":7249},[],[7251],{"type":70,"value":7252},"Robotics-GR00T-GR1",{"type":70,"value":7254}," (DreamGen ref) \u002F ",{"type":64,"tag":86,"props":7256,"children":7258},{"className":7257},[],[7259],{"type":70,"value":7260},"-Eval",{"type":64,"tag":1986,"props":7262,"children":7263},{},[7264],{"type":70,"value":6974},{"type":64,"tag":1986,"props":7266,"children":7267},{},[7268],{"type":70,"value":7269},"reference assets only",{"type":64,"tag":1964,"props":7271,"children":7272},{},[7273,7282,7286],{"type":64,"tag":1986,"props":7274,"children":7275},{},[7276],{"type":64,"tag":86,"props":7277,"children":7279},{"className":7278},[],[7280],{"type":70,"value":7281},"Robotics-mindmap-*",{"type":64,"tag":1986,"props":7283,"children":7284},{},[7285],{"type":70,"value":6974},{"type":64,"tag":1986,"props":7287,"children":7288},{},[7289,7298],{"type":64,"tag":206,"props":7290,"children":7292},{"href":5160,"rel":7291},[209],[7293],{"type":64,"tag":86,"props":7294,"children":7296},{"className":7295},[],[7297],{"type":70,"value":4888},{"type":70,"value":7299},", Isaac Lab",{"type":64,"tag":1964,"props":7301,"children":7302},{},[7303,7312,7323],{"type":64,"tag":1986,"props":7304,"children":7305},{},[7306],{"type":64,"tag":86,"props":7307,"children":7309},{"className":7308},[],[7310],{"type":70,"value":7311},"Robotics-NuRec",{"type":64,"tag":1986,"props":7313,"children":7314},{},[7315],{"type":64,"tag":206,"props":7316,"children":7317},{"href":3037},[7318],{"type":64,"tag":86,"props":7319,"children":7321},{"className":7320},[],[7322],{"type":70,"value":3037},{"type":64,"tag":1986,"props":7324,"children":7325},{},[7326,7328],{"type":70,"value":7327},"Isaac Sim 5.1, ",{"type":64,"tag":206,"props":7329,"children":7331},{"href":5333,"rel":7330},[209],[7332],{"type":64,"tag":86,"props":7333,"children":7335},{"className":7334},[],[7336],{"type":70,"value":5341},{"type":64,"tag":1964,"props":7338,"children":7339},{},[7340,7349,7362],{"type":64,"tag":1986,"props":7341,"children":7342},{},[7343],{"type":64,"tag":86,"props":7344,"children":7346},{"className":7345},[],[7347],{"type":70,"value":7348},"NuRec-PPISP",{"type":64,"tag":1986,"props":7350,"children":7351},{},[7352,7360],{"type":64,"tag":206,"props":7353,"children":7354},{"href":3037},[7355],{"type":64,"tag":86,"props":7356,"children":7358},{"className":7357},[],[7359],{"type":70,"value":3037},{"type":70,"value":7361}," (3DGRUT \u002F GSplat benchmarking)",{"type":64,"tag":1986,"props":7363,"children":7364},{},[7365],{"type":70,"value":6974},{"type":64,"tag":1964,"props":7367,"children":7368},{},[7369,7378,7382],{"type":64,"tag":1986,"props":7370,"children":7371},{},[7372],{"type":64,"tag":86,"props":7373,"children":7375},{"className":7374},[],[7376],{"type":70,"value":7377},"Robotics-Open-H-Embodiment",{"type":64,"tag":1986,"props":7379,"children":7380},{},[7381],{"type":70,"value":6974},{"type":64,"tag":1986,"props":7383,"children":7384},{},[7385,7390,7391],{"type":64,"tag":86,"props":7386,"children":7388},{"className":7387},[],[7389],{"type":70,"value":5589},{"type":70,"value":101},{"type":64,"tag":86,"props":7392,"children":7394},{"className":7393},[],[7395],{"type":70,"value":5596},{"type":64,"tag":1964,"props":7397,"children":7398},{},[7399,7408,7412],{"type":64,"tag":1986,"props":7400,"children":7401},{},[7402],{"type":64,"tag":86,"props":7403,"children":7405},{"className":7404},[],[7406],{"type":70,"value":7407},"Robotics-GraspGen",{"type":64,"tag":1986,"props":7409,"children":7410},{},[7411],{"type":70,"value":6974},{"type":64,"tag":1986,"props":7413,"children":7414},{},[7415],{"type":70,"value":7416},"Ships its own visualisation scripts; Isaac Sim for replay",{"type":64,"tag":1964,"props":7418,"children":7419},{},[7420,7435,7439],{"type":64,"tag":1986,"props":7421,"children":7422},{},[7423,7429,7430],{"type":64,"tag":86,"props":7424,"children":7426},{"className":7425},[],[7427],{"type":70,"value":7428},"Robotics-PhysicalAssets-VoMP",{"type":70,"value":6473},{"type":64,"tag":86,"props":7431,"children":7433},{"className":7432},[],[7434],{"type":70,"value":7260},{"type":64,"tag":1986,"props":7436,"children":7437},{},[7438],{"type":70,"value":6974},{"type":64,"tag":1986,"props":7440,"children":7441},{},[7442],{"type":70,"value":7443},"VoMP model",{"type":64,"tag":1964,"props":7445,"children":7446},{},[7447,7456,7460],{"type":64,"tag":1986,"props":7448,"children":7449},{},[7450],{"type":64,"tag":86,"props":7451,"children":7453},{"className":7452},[],[7454],{"type":70,"value":7455},"SimReady-Warehouse-01",{"type":64,"tag":1986,"props":7457,"children":7458},{},[7459],{"type":70,"value":6974},{"type":64,"tag":1986,"props":7461,"children":7462},{},[7463],{"type":70,"value":7464},"Isaac Sim 4.x, Omniverse SDG",{"type":64,"tag":1964,"props":7466,"children":7467},{},[7468,7477,7481],{"type":64,"tag":1986,"props":7469,"children":7470},{},[7471],{"type":64,"tag":86,"props":7472,"children":7474},{"className":7473},[],[7475],{"type":70,"value":7476},"DigitalCousin-Assets",{"type":64,"tag":1986,"props":7478,"children":7479},{},[7480],{"type":70,"value":6974},{"type":64,"tag":1986,"props":7482,"children":7483},{},[7484],{"type":70,"value":7485},"Isaac Sim",{"type":64,"tag":1964,"props":7487,"children":7488},{},[7489,7498,7502],{"type":64,"tag":1986,"props":7490,"children":7491},{},[7492],{"type":64,"tag":86,"props":7493,"children":7495},{"className":7494},[],[7496],{"type":70,"value":7497},"SmartSpaces",{"type":64,"tag":1986,"props":7499,"children":7500},{},[7501],{"type":70,"value":6974},{"type":64,"tag":1986,"props":7503,"children":7504},{},[7505],{"type":70,"value":7506},"AI City Challenge eval server",{"type":64,"tag":1964,"props":7508,"children":7509},{},[7510,7518,7522],{"type":64,"tag":1986,"props":7511,"children":7512},{},[7513],{"type":64,"tag":86,"props":7514,"children":7516},{"className":7515},[],[7517],{"type":70,"value":1564},{"type":64,"tag":1986,"props":7519,"children":7520},{},[7521],{"type":70,"value":6974},{"type":64,"tag":1986,"props":7523,"children":7524},{},[7525],{"type":70,"value":7526},"Warehouse VLM benchmark (no upstream skill)",{"type":64,"tag":1964,"props":7528,"children":7529},{},[7530,7539,7543],{"type":64,"tag":1986,"props":7531,"children":7532},{},[7533],{"type":64,"tag":86,"props":7534,"children":7536},{"className":7535},[],[7537],{"type":70,"value":7538},"SpatialIntelligence-Lyra-SDG",{"type":64,"tag":1986,"props":7540,"children":7541},{},[7542],{"type":70,"value":6974},{"type":64,"tag":1986,"props":7544,"children":7545},{},[7546,7555,7556,7561],{"type":64,"tag":206,"props":7547,"children":7549},{"href":6644,"rel":7548},[209],[7550],{"type":64,"tag":86,"props":7551,"children":7553},{"className":7552},[],[7554],{"type":70,"value":6504},{"type":70,"value":101},{"type":64,"tag":86,"props":7557,"children":7559},{"className":7558},[],[7560],{"type":70,"value":3237},{"type":70,"value":7562}," (GEN3C lineage)",{"type":64,"tag":73,"props":7564,"children":7566},{"id":7565},"verify",[7567],{"type":70,"value":7568},"Verify",{"type":64,"tag":80,"props":7570,"children":7571},{},[7572],{"type":70,"value":7573},"After downloading any dataset:",{"type":64,"tag":351,"props":7575,"children":7577},{"className":353,"code":7576,"language":45,"meta":355,"style":355},"ls \u003Clocal-dir>\ndu -sh \u003Clocal-dir>          # confirm size matches the table above (within 10 %)\n\n# For LeRobot v2.x:\npython -c \"from lerobot.common.datasets.lerobot_dataset import LeRobotDataset; \\\n           d = LeRobotDataset('\u003Clocal-dir>'); print(d.meta.info)\"\n\n# For NCore V4 (.zarr.itar):\nncore_vis \u003Clocal-dir>\u002Fclips\u002F\u003Cuuid>\u002Fpai_\u003Cuuid>.json   # via ..\u002Fncore\u002FSKILL.md\n\n# For USDZ scenes:\n# load in Isaac Sim 5.1 — File → Open → \u003Cscene>.usdz\n",[7578],{"type":64,"tag":86,"props":7579,"children":7580},{"__ignoreMap":355},[7581,7606,7640,7647,7655,7681,7693,7700,7708,7786,7793,7801],{"type":64,"tag":361,"props":7582,"children":7583},{"class":363,"line":364},[7584,7589,7593,7598,7602],{"type":64,"tag":361,"props":7585,"children":7586},{"style":499},[7587],{"type":70,"value":7588},"ls",{"type":64,"tag":361,"props":7590,"children":7591},{"style":384},[7592],{"type":70,"value":1248},{"type":64,"tag":361,"props":7594,"children":7595},{"style":390},[7596],{"type":70,"value":7597},"local-di",{"type":64,"tag":361,"props":7599,"children":7600},{"style":401},[7601],{"type":70,"value":1314},{"type":64,"tag":361,"props":7603,"children":7604},{"style":384},[7605],{"type":70,"value":1080},{"type":64,"tag":361,"props":7607,"children":7608},{"class":363,"line":374},[7609,7614,7619,7623,7627,7631,7635],{"type":64,"tag":361,"props":7610,"children":7611},{"style":499},[7612],{"type":70,"value":7613},"du",{"type":64,"tag":361,"props":7615,"children":7616},{"style":390},[7617],{"type":70,"value":7618}," -sh",{"type":64,"tag":361,"props":7620,"children":7621},{"style":384},[7622],{"type":70,"value":1248},{"type":64,"tag":361,"props":7624,"children":7625},{"style":390},[7626],{"type":70,"value":7597},{"type":64,"tag":361,"props":7628,"children":7629},{"style":401},[7630],{"type":70,"value":1314},{"type":64,"tag":361,"props":7632,"children":7633},{"style":384},[7634],{"type":70,"value":1043},{"type":64,"tag":361,"props":7636,"children":7637},{"style":368},[7638],{"type":70,"value":7639},"          # confirm size matches the table above (within 10 %)\n",{"type":64,"tag":361,"props":7641,"children":7642},{"class":363,"line":886},[7643],{"type":64,"tag":361,"props":7644,"children":7645},{"emptyLinePlaceholder":890},[7646],{"type":70,"value":893},{"type":64,"tag":361,"props":7648,"children":7649},{"class":363,"line":30},[7650],{"type":64,"tag":361,"props":7651,"children":7652},{"style":368},[7653],{"type":70,"value":7654},"# For LeRobot v2.x:\n",{"type":64,"tag":361,"props":7656,"children":7657},{"class":363,"line":935},[7658,7662,7667,7671,7676],{"type":64,"tag":361,"props":7659,"children":7660},{"style":499},[7661],{"type":70,"value":46},{"type":64,"tag":361,"props":7663,"children":7664},{"style":390},[7665],{"type":70,"value":7666}," -c",{"type":64,"tag":361,"props":7668,"children":7669},{"style":384},[7670],{"type":70,"value":387},{"type":64,"tag":361,"props":7672,"children":7673},{"style":390},[7674],{"type":70,"value":7675},"from lerobot.common.datasets.lerobot_dataset import LeRobotDataset; ",{"type":64,"tag":361,"props":7677,"children":7678},{"style":401},[7679],{"type":70,"value":7680},"\\\n",{"type":64,"tag":361,"props":7682,"children":7683},{"class":363,"line":1119},[7684,7689],{"type":64,"tag":361,"props":7685,"children":7686},{"style":390},[7687],{"type":70,"value":7688},"           d = LeRobotDataset('\u003Clocal-dir>'); print(d.meta.info)",{"type":64,"tag":361,"props":7690,"children":7691},{"style":384},[7692],{"type":70,"value":609},{"type":64,"tag":361,"props":7694,"children":7695},{"class":363,"line":1136},[7696],{"type":64,"tag":361,"props":7697,"children":7698},{"emptyLinePlaceholder":890},[7699],{"type":70,"value":893},{"type":64,"tag":361,"props":7701,"children":7702},{"class":363,"line":1162},[7703],{"type":64,"tag":361,"props":7704,"children":7705},{"style":368},[7706],{"type":70,"value":7707},"# For NCore V4 (.zarr.itar):\n",{"type":64,"tag":361,"props":7709,"children":7710},{"class":363,"line":1176},[7711,7716,7720,7724,7728,7732,7737,7741,7746,7751,7755,7760,7764,7768,7772,7776,7781],{"type":64,"tag":361,"props":7712,"children":7713},{"style":499},[7714],{"type":70,"value":7715},"ncore_vis",{"type":64,"tag":361,"props":7717,"children":7718},{"style":384},[7719],{"type":70,"value":1248},{"type":64,"tag":361,"props":7721,"children":7722},{"style":390},[7723],{"type":70,"value":7597},{"type":64,"tag":361,"props":7725,"children":7726},{"style":401},[7727],{"type":70,"value":1314},{"type":64,"tag":361,"props":7729,"children":7730},{"style":384},[7731],{"type":70,"value":1043},{"type":64,"tag":361,"props":7733,"children":7734},{"style":390},[7735],{"type":70,"value":7736},"\u002Fclips\u002F",{"type":64,"tag":361,"props":7738,"children":7739},{"style":384},[7740],{"type":70,"value":1028},{"type":64,"tag":361,"props":7742,"children":7743},{"style":390},[7744],{"type":70,"value":7745},"uui",{"type":64,"tag":361,"props":7747,"children":7748},{"style":401},[7749],{"type":70,"value":7750},"d",{"type":64,"tag":361,"props":7752,"children":7753},{"style":384},[7754],{"type":70,"value":1043},{"type":64,"tag":361,"props":7756,"children":7757},{"style":390},[7758],{"type":70,"value":7759},"\u002Fpai_",{"type":64,"tag":361,"props":7761,"children":7762},{"style":384},[7763],{"type":70,"value":1028},{"type":64,"tag":361,"props":7765,"children":7766},{"style":390},[7767],{"type":70,"value":7745},{"type":64,"tag":361,"props":7769,"children":7770},{"style":401},[7771],{"type":70,"value":7750},{"type":64,"tag":361,"props":7773,"children":7774},{"style":384},[7775],{"type":70,"value":1043},{"type":64,"tag":361,"props":7777,"children":7778},{"style":390},[7779],{"type":70,"value":7780},".json",{"type":64,"tag":361,"props":7782,"children":7783},{"style":368},[7784],{"type":70,"value":7785},"   # via ..\u002Fncore\u002FSKILL.md\n",{"type":64,"tag":361,"props":7787,"children":7788},{"class":363,"line":1184},[7789],{"type":64,"tag":361,"props":7790,"children":7791},{"emptyLinePlaceholder":890},[7792],{"type":70,"value":893},{"type":64,"tag":361,"props":7794,"children":7795},{"class":363,"line":1193},[7796],{"type":64,"tag":361,"props":7797,"children":7798},{"style":368},[7799],{"type":70,"value":7800},"# For USDZ scenes:\n",{"type":64,"tag":361,"props":7802,"children":7803},{"class":363,"line":1237},[7804],{"type":64,"tag":361,"props":7805,"children":7806},{"style":368},[7807],{"type":70,"value":7808},"# load in Isaac Sim 5.1 — File → Open → \u003Cscene>.usdz\n",{"type":64,"tag":80,"props":7810,"children":7811},{},[7812],{"type":70,"value":7813},"GREEN when:",{"type":64,"tag":152,"props":7815,"children":7816},{},[7817,7822,7835],{"type":64,"tag":156,"props":7818,"children":7819},{},[7820],{"type":70,"value":7821},"File counts and total size match the dataset card to within ~10 %.",{"type":64,"tag":156,"props":7823,"children":7824},{},[7825,7827,7833],{"type":70,"value":7826},"For gated datasets, the download did not silently terminate at the\nlicense-agreement page (re-",{"type":64,"tag":86,"props":7828,"children":7830},{"className":7829},[],[7831],{"type":70,"value":7832},"hf auth login",{"type":70,"value":7834}," if the first chunk is HTML).",{"type":64,"tag":156,"props":7836,"children":7837},{},[7838,7840,7846],{"type":70,"value":7839},"For LeRobot datasets, ",{"type":64,"tag":86,"props":7841,"children":7843},{"className":7842},[],[7844],{"type":70,"value":7845},"meta\u002Finfo.json",{"type":70,"value":7847}," parses and the episode count\nmatches the card.",{"type":64,"tag":73,"props":7849,"children":7851},{"id":7850},"troubleshooting",[7852],{"type":70,"value":7853},"Troubleshooting",{"type":64,"tag":152,"props":7855,"children":7856},{},[7857,7885,7895,7925,7964,7996,8014,8032,8049],{"type":64,"tag":156,"props":7858,"children":7859},{},[7860,7869,7871,7876,7878,7883],{"type":64,"tag":129,"props":7861,"children":7862},{},[7863],{"type":64,"tag":86,"props":7864,"children":7866},{"className":7865},[],[7867],{"type":70,"value":7868},"Repo gated. Cannot access … 401",{"type":70,"value":7870}," — open the dataset URL in a\nbrowser, click ",{"type":64,"tag":1405,"props":7872,"children":7873},{},[7874],{"type":70,"value":7875},"Agree",{"type":70,"value":7877}," on the license \u002F contact-info form, ",{"type":64,"tag":129,"props":7879,"children":7880},{},[7881],{"type":70,"value":7882},"then",{"type":70,"value":7884}," retry\nwith the same token. Tokens don't get auto-refreshed when a new\nagreement appears (re-accept after major version bumps).",{"type":64,"tag":156,"props":7886,"children":7887},{},[7888,7893],{"type":64,"tag":129,"props":7889,"children":7890},{},[7891],{"type":70,"value":7892},"First chunk is 5 KB of HTML",{"type":70,"value":7894}," — same cause as above.",{"type":64,"tag":156,"props":7896,"children":7897},{},[7898,7903,7905,7910,7912,7917,7919,7923],{"type":64,"tag":129,"props":7899,"children":7900},{},[7901],{"type":70,"value":7902},"Download hangs indefinitely on AV (133 TB)",{"type":70,"value":7904}," — you almost certainly\ndon't want the whole thing. Use ",{"type":64,"tag":86,"props":7906,"children":7908},{"className":7907},[],[7909],{"type":70,"value":316},{"type":70,"value":7911}," to filter\n",{"type":64,"tag":86,"props":7913,"children":7915},{"className":7914},[],[7916],{"type":70,"value":1384},{"type":70,"value":7918}," (default — see\n",{"type":64,"tag":206,"props":7920,"children":7921},{"href":648},[7922],{"type":70,"value":651},{"type":70,"value":7924},"), plus\nany sensor \u002F country \u002F split mask, BEFORE pulling.",{"type":64,"tag":156,"props":7926,"children":7927},{},[7928,7933,7935,7941,7943,7948,7950,7955,7957,7962],{"type":64,"tag":129,"props":7929,"children":7930},{},[7931],{"type":70,"value":7932},"AV clip downloaded but downstream NCore \u002F NuRec \u002F Asset-Harvester\nrejects it",{"type":70,"value":7934}," — double-check ",{"type":64,"tag":86,"props":7936,"children":7938},{"className":7937},[],[7939],{"type":70,"value":7940},"data_collection['platform_class']",{"type":70,"value":7942}," for\nthat clip; only ",{"type":64,"tag":86,"props":7944,"children":7946},{"className":7945},[],[7947],{"type":70,"value":659},{"type":70,"value":7949}," is validated. If it's ",{"type":64,"tag":86,"props":7951,"children":7953},{"className":7952},[],[7954],{"type":70,"value":1394},{"type":70,"value":7956},",\neither swap to a ",{"type":64,"tag":86,"props":7958,"children":7960},{"className":7959},[],[7961],{"type":70,"value":659},{"type":70,"value":7963}," clip or accept that the downstream\nreconstruction tooling will not work.",{"type":64,"tag":156,"props":7965,"children":7966},{},[7967,7972,7974,7980,7982,7988,7990,7995],{"type":64,"tag":129,"props":7968,"children":7969},{},[7970],{"type":70,"value":7971},"Cosmos-Drive-Dreams 3 TB on a small disk",{"type":70,"value":7973}," — pass\n",{"type":64,"tag":86,"props":7975,"children":7977},{"className":7976},[],[7978],{"type":70,"value":7979},"--file_types synthetic",{"type":70,"value":7981}," (700 GB) or ",{"type":64,"tag":86,"props":7983,"children":7985},{"className":7984},[],[7986],{"type":70,"value":7987},"--file_types hdmap",{"type":70,"value":7989}," (small) to\nthe official ",{"type":64,"tag":86,"props":7991,"children":7993},{"className":7992},[],[7994],{"type":70,"value":1520},{"type":70,"value":621},{"type":64,"tag":156,"props":7997,"children":7998},{},[7999,8004,8006,8012],{"type":64,"tag":129,"props":8000,"children":8001},{},[8002],{"type":70,"value":8003},"Lyra-SDG 25 TB out of disk",{"type":70,"value":8005}," — ",{"type":64,"tag":86,"props":8007,"children":8009},{"className":8008},[],[8010],{"type":70,"value":8011},"hf download --include \"tar\u002Fstatic_*\" --exclude \"tar\u002Fdynamic_*\"",{"type":70,"value":8013}," to take just the 3D half (or vice-versa).",{"type":64,"tag":156,"props":8015,"children":8016},{},[8017,8022,8024,8030],{"type":64,"tag":129,"props":8018,"children":8019},{},[8020],{"type":70,"value":8021},"Cosmos-Synthetic looks empty (2.59 kB)",{"type":70,"value":8023}," — it's a pointer page\nredirecting to ",{"type":64,"tag":86,"props":8025,"children":8027},{"className":8026},[],[8028],{"type":70,"value":8029},"Cosmos-Drive-Dreams",{"type":70,"value":8031},". Use that.",{"type":64,"tag":156,"props":8033,"children":8034},{},[8035,8040,8042,8047],{"type":64,"tag":129,"props":8036,"children":8037},{},[8038],{"type":70,"value":8039},"mindmap dataset only has 10 demos when the card says 200 \u002F 1000",{"type":70,"value":8041}," —\nexpected; regenerate the rest from the HDF5 via the upstream\n",{"type":64,"tag":86,"props":8043,"children":8045},{"className":8044},[],[8046],{"type":70,"value":4997},{"type":70,"value":8048}," data-generation docs.",{"type":64,"tag":156,"props":8050,"children":8051},{},[8052,8057],{"type":64,"tag":129,"props":8053,"children":8054},{},[8055],{"type":70,"value":8056},"AV license expired (12 months)",{"type":70,"value":8058}," — re-accept on the HF page; this\nis by-design per the License Agreement section 7.",{"type":64,"tag":73,"props":8060,"children":8062},{"id":8061},"limitations",[8063],{"type":70,"value":8064},"Limitations",{"type":64,"tag":152,"props":8066,"children":8067},{},[8068,8078,8102,8148,8172,8197],{"type":64,"tag":156,"props":8069,"children":8070},{},[8071,8076],{"type":64,"tag":129,"props":8072,"children":8073},{},[8074],{"type":70,"value":8075},"Catalogue-only.",{"type":70,"value":8077}," This skill does not actually run any\npipeline — it routes the agent to a sibling skill once the right\ndataset is chosen.",{"type":64,"tag":156,"props":8079,"children":8080},{},[8081,8086,8088,8093,8095,8100],{"type":64,"tag":129,"props":8082,"children":8083},{},[8084],{"type":70,"value":8085},"NVIDIA-only.",{"type":70,"value":8087}," Non-NVIDIA datasets (Waymo, nuScenes, KITTI,\nPandaSet, …) are out of scope; the ",{"type":64,"tag":86,"props":8089,"children":8091},{"className":8090},[],[8092],{"type":70,"value":99},{"type":70,"value":8094}," skill covers how to\nconvert ",{"type":64,"tag":1405,"props":8096,"children":8097},{},[8098],{"type":70,"value":8099},"those",{"type":70,"value":8101}," into NCore V4 manually.",{"type":64,"tag":156,"props":8103,"children":8104},{},[8105,8110,8112,8118,8119,8125,8127,8132,8134,8139,8141,8146],{"type":64,"tag":129,"props":8106,"children":8107},{},[8108],{"type":70,"value":8109},"Gating is enforced by Hugging Face, not by this skill.",{"type":70,"value":8111}," Several\ndatasets (e.g. ",{"type":64,"tag":86,"props":8113,"children":8115},{"className":8114},[],[8116],{"type":70,"value":8117},"PhysicalAI-Autonomous-Vehicles*",{"type":70,"value":109},{"type":64,"tag":86,"props":8120,"children":8122},{"className":8121},[],[8123],{"type":70,"value":8124},"nvidia\u002FFixer",{"type":70,"value":8126},"-adjacent assets, mindmap large variants) require\n",{"type":64,"tag":1405,"props":8128,"children":8129},{},[8130],{"type":70,"value":8131},"both",{"type":70,"value":8133}," license acceptance on the HF page ",{"type":64,"tag":1405,"props":8135,"children":8136},{},[8137],{"type":70,"value":8138},"and",{"type":70,"value":8140}," a valid ",{"type":64,"tag":86,"props":8142,"children":8144},{"className":8143},[],[8145],{"type":70,"value":261},{"type":70,"value":8147},".\nAcceptances expire (AV is annual) and must be re-clicked.",{"type":64,"tag":156,"props":8149,"children":8150},{},[8151,8156,8158,8163,8164,8170],{"type":64,"tag":129,"props":8152,"children":8153},{},[8154],{"type":70,"value":8155},"Footprint can be enormous.",{"type":70,"value":8157}," PhysicalAI-Autonomous-Vehicles is\n~133 TB; Lyra-SDG ~25 TB; Cosmos-Drive-Dreams ~3 TB. Always check\nfree disk and use ",{"type":64,"tag":86,"props":8159,"children":8161},{"className":8160},[],[8162],{"type":70,"value":308},{"type":70,"value":6473},{"type":64,"tag":86,"props":8165,"children":8167},{"className":8166},[],[8168],{"type":70,"value":8169},"--exclude",{"type":70,"value":8171}," filters before kicking\noff a full pull.",{"type":64,"tag":156,"props":8173,"children":8174},{},[8175,8180,8182,8187,8189,8195],{"type":64,"tag":129,"props":8176,"children":8177},{},[8178],{"type":70,"value":8179},"The catalogue drifts.",{"type":70,"value":8181}," NVIDIA ships new ",{"type":64,"tag":86,"props":8183,"children":8185},{"className":8184},[],[8186],{"type":70,"value":91},{"type":70,"value":8188}," datasets\nregularly. If a dataset isn't in this skill, browse\n",{"type":64,"tag":206,"props":8190,"children":8193},{"href":8191,"rel":8192},"https:\u002F\u002Fhuggingface.co\u002Fnvidia?search_models=PhysicalAI-",[209],[8194],{"type":70,"value":8191},{"type":70,"value":8196}," and\nconsider opening a PR to update Section 2 \u002F the relevant family.",{"type":64,"tag":156,"props":8198,"children":8199},{},[8200,8205],{"type":64,"tag":129,"props":8201,"children":8202},{},[8203],{"type":70,"value":8204},"No GPU \u002F runtime checks here.",{"type":70,"value":8206}," Storage and HF auth are the only\nprerequisites this skill validates; per-dataset compute needs\nbelong to the consuming sibling skill.",{"type":64,"tag":73,"props":8208,"children":8210},{"id":8209},"source-links",[8211],{"type":70,"value":8212},"Source links",{"type":64,"tag":152,"props":8214,"children":8215},{},[8216,8233,8243,8253,8263,8274,8284,8294,8304,8314,8324],{"type":64,"tag":156,"props":8217,"children":8218},{},[8219,8221,8226,8228],{"type":70,"value":8220},"HF org (filter ",{"type":64,"tag":86,"props":8222,"children":8224},{"className":8223},[],[8225],{"type":70,"value":219},{"type":70,"value":8227},"): ",{"type":64,"tag":206,"props":8229,"children":8231},{"href":56,"rel":8230},[209],[8232],{"type":70,"value":56},{"type":64,"tag":156,"props":8234,"children":8235},{},[8236,8238],{"type":70,"value":8237},"Curated collection (29 items): ",{"type":64,"tag":206,"props":8239,"children":8241},{"href":57,"rel":8240},[209],[8242],{"type":70,"value":57},{"type":64,"tag":156,"props":8244,"children":8245},{},[8246,8248],{"type":70,"value":8247},"Cosmos Dataset Search (semantic preview of AV): ",{"type":64,"tag":206,"props":8249,"children":8251},{"href":58,"rel":8250},[209],[8252],{"type":70,"value":58},{"type":64,"tag":156,"props":8254,"children":8255},{},[8256,8258],{"type":70,"value":8257},"AV Python toolkit: ",{"type":64,"tag":206,"props":8259,"children":8261},{"href":59,"rel":8260},[209],[8262],{"type":70,"value":59},{"type":64,"tag":156,"props":8264,"children":8265},{},[8266,8268],{"type":70,"value":8267},"NCore: ",{"type":64,"tag":206,"props":8269,"children":8272},{"href":8270,"rel":8271},"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fncore",[209],[8273],{"type":70,"value":8270},{"type":64,"tag":156,"props":8275,"children":8276},{},[8277,8279],{"type":70,"value":8278},"Cosmos-Drive-Dreams downloader: ",{"type":64,"tag":206,"props":8280,"children":8282},{"href":1524,"rel":8281},[209],[8283],{"type":70,"value":1524},{"type":64,"tag":156,"props":8285,"children":8286},{},[8287,8289],{"type":70,"value":8288},"Lyra repo: ",{"type":64,"tag":206,"props":8290,"children":8292},{"href":6644,"rel":8291},[209],[8293],{"type":70,"value":6644},{"type":64,"tag":156,"props":8295,"children":8296},{},[8297,8299],{"type":70,"value":8298},"mindmap repo: ",{"type":64,"tag":206,"props":8300,"children":8302},{"href":5160,"rel":8301},[209],[8303],{"type":70,"value":5160},{"type":64,"tag":156,"props":8305,"children":8306},{},[8307,8309],{"type":70,"value":8308},"Isaac-GR00T repo: ",{"type":64,"tag":206,"props":8310,"children":8312},{"href":7186,"rel":8311},[209],[8313],{"type":70,"value":7186},{"type":64,"tag":156,"props":8315,"children":8316},{},[8317,8319],{"type":70,"value":8318},"DreamDojo project: ",{"type":64,"tag":206,"props":8320,"children":8322},{"href":4669,"rel":8321},[209],[8323],{"type":70,"value":4669},{"type":64,"tag":156,"props":8325,"children":8326},{},[8327,8329],{"type":70,"value":8328},"AI City Challenge eval: ",{"type":64,"tag":206,"props":8330,"children":8333},{"href":8331,"rel":8332},"https:\u002F\u002Feval.aicitychallenge.org\u002F",[209],[8334],{"type":70,"value":8331},{"type":64,"tag":8336,"props":8337,"children":8338},"style",{},[8339],{"type":70,"value":8340},"html .light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html.light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"items":8342,"total":1119},[8343,8359,8371,8381,8391,8401],{"slug":115,"name":115,"fn":8344,"description":8345,"org":8346,"tags":8347,"stars":26,"repoUrl":27,"updatedAt":8358},"extract 3D Gaussian assets with Asset Harvester","Use to install and run NVIDIA Asset Harvester (Apache-2.0) to extract per-object 3D Gaussian Splat assets (`gaussians.ply`) from AV NCore V4 clips or masked single images via SparseViewDiT + TokenGS, optionally producing `metadata.yaml` for NuRec object insertion. Do NOT use for full-scene reconstruction (use `nre`) or for inputs without per-object masks.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8348,8351,8354,8355],{"name":8349,"slug":8350,"type":15},"3D","3d",{"name":8352,"slug":8353,"type":15},"Computer Vision","computer-vision",{"name":9,"slug":8,"type":15},{"name":8356,"slug":8357,"type":15},"Simulation","simulation","2026-07-14T05:32:37.216976",{"slug":99,"name":99,"fn":8360,"description":8361,"org":8362,"tags":8363,"stars":26,"repoUrl":27,"updatedAt":8370},"convert sensor datasets to NVIDIA NCore format","Use when converting any sensor dataset into NVIDIA NCore V4 format (and feeding it to NuRec or a robotics-to-sim \"r2s\" pipeline). Covers ingesting raw cameras, LiDARs, radars, IMUs, depth or stereo into V4 sequences; authoring a new converter from the template; adapting PAI \u002F Waymo \u002F PandaSet \u002F NuScenes to V4; handling non-AV rigs (mono+depth, mono+lidar, stereo, multi-stereo, RGB-D, COLMAP \u002F SfM, ROS2 bag); and diagnosing a broken converter against `validate.py`. Do NOT use to train reconstructions (use `nre`) or to extract per-object 3D assets (use `asset-harvester`). Trigger keywords: ncore, ncore v4, convert, ingest, zarr, itar, nurec, waymo, pandaset, nuscenes, pai, hyperion, colmap, scannetpp, stereo, multi-stereo, mono+depth, mono+lidar, kitti, sfm, camera, lidar, radar, imu, cuboid, poses, intrinsics, ego mask, ros2, rosbag, mcap, realsense, zed, rgb-d, r2s, robotics, sam2.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8364,8367,8368,8369],{"name":8365,"slug":8366,"type":15},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":15},{"name":13,"slug":14,"type":15},{"name":8356,"slug":8357,"type":15},"2026-07-14T05:32:35.853952",{"slug":107,"name":107,"fn":8372,"description":8373,"org":8374,"tags":8375,"stars":26,"repoUrl":27,"updatedAt":8380},"train 3D Gaussian Splat models with NRE","Use to drive NVIDIA Omniverse NuRec \u002F Neural Reconstruction Engine (NRE) via the public NGC containers nvcr.io\u002Fnvidia\u002Fnre\u002Fnre and nvcr.io\u002Fnvidia\u002Fnre\u002Fnre-tools (NGC_API_KEY required) — train 3DGUT Gaussian reconstructions from NCore clips, generate aux data, render frames or LiDAR sweeps (local or warm `serve-grpc`), export PLY\u002Fdepth\u002Fmesh\u002FUSDZ, edit actors, and evaluate metrics. Do NOT use for per-object asset capture (use `asset-harvester`) or sensor-to-NCore conversion (use `ncore`).",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8376,8377,8378,8379],{"name":8349,"slug":8350,"type":15},{"name":23,"slug":24,"type":15},{"name":9,"slug":8,"type":15},{"name":8356,"slug":8357,"type":15},"2026-07-14T05:32:34.548655",{"slug":122,"name":122,"fn":8382,"description":8383,"org":8384,"tags":8385,"stars":26,"repoUrl":27,"updatedAt":8390},"harmonize and fine-tune 3D reconstruction frames","Use to run NVIDIA DiffusionHarmonizer (public successor to the older Fixer recipes) to enhance, harmonize, evaluate, or fine-tune novel-view frames from NRE \u002F NuRec \u002F 3DGS \u002F NeRF reconstructions. Do NOT use for training the 3D reconstruction itself (use `nre`) or for sensor-to-NCore conversion (use `ncore`).",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8386,8387,8388,8389],{"name":8349,"slug":8350,"type":15},{"name":23,"slug":24,"type":15},{"name":9,"slug":8,"type":15},{"name":8356,"slug":8357,"type":15},"2026-07-14T05:32:38.465728",{"slug":8392,"name":8392,"fn":8393,"description":8394,"org":8395,"tags":8396,"stars":26,"repoUrl":27,"updatedAt":8400},"nurec-index","route NVIDIA NuRec and asset tasks","Router for NVIDIA NuRec \u002F NRE \u002F 3DGUT \u002F USDZ \u002F NCore V4 \u002F asset harvest \u002F frame cleanup tasks — picks the right sibling (nre, ncore, asset-harvester, nurec-fixer, physical-ai-datasets). Use when the sub-skill is unclear or a multi-stage pipeline is needed; do NOT use for non-NuRec tasks or to run any pipeline itself.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8397,8398,8399],{"name":8349,"slug":8350,"type":15},{"name":9,"slug":8,"type":15},{"name":8356,"slug":8357,"type":15},"2026-07-14T05:32:31.902229",{"slug":4,"name":4,"fn":5,"description":6,"org":8402,"tags":8403,"stars":26,"repoUrl":27,"updatedAt":28},{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8404,8405,8406,8407,8408],{"name":17,"slug":18,"type":15},{"name":20,"slug":21,"type":15},{"name":23,"slug":24,"type":15},{"name":9,"slug":8,"type":15},{"name":13,"slug":14,"type":15},{"items":8410,"total":8566},[8411,8429,8446,8457,8469,8483,8496,8508,8521,8532,8546,8555],{"slug":8412,"name":8412,"fn":8413,"description":8414,"org":8415,"tags":8416,"stars":8426,"repoUrl":8427,"updatedAt":8428},"nemoclaw-user-guide","retrieve NemoClaw documentation and configuration","Guides human users' AI agents to the NemoClaw docs MCP server and canonical Fern documentation in Markdown form. Use when users ask how to install, configure, operate, troubleshoot, secure, or learn NemoClaw with an AI coding assistant. Trigger keywords - nemoclaw docs, use nemoclaw with ai agent, nemoclaw mcp docs, nemoclaw install help, nemoclaw quickstart, nemoclaw markdown docs, llms.txt, agent skills.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8417,8420,8423],{"name":8418,"slug":8419,"type":15},"Documentation","documentation",{"name":8421,"slug":8422,"type":15},"MCP","mcp",{"name":8424,"slug":8425,"type":15},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":8430,"name":8430,"fn":8431,"description":8432,"org":8433,"tags":8434,"stars":8443,"repoUrl":8444,"updatedAt":8445},"mcore-build-and-dependency","manage Megatron-LM development environments","Container-based dev environment setup and dependency management for Megatron-LM. Covers acquiring and launching the CI container, uv package management, and updating uv.lock.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8435,8438,8441],{"name":8436,"slug":8437,"type":15},"Containers","containers",{"name":8439,"slug":8440,"type":15},"Deployment","deployment",{"name":8442,"slug":46,"type":15},"Python",17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":8447,"name":8447,"fn":8448,"description":8449,"org":8450,"tags":8451,"stars":8443,"repoUrl":8444,"updatedAt":8456},"mcore-bump-base-image","update NVIDIA PyTorch base images","Bump the NVIDIA PyTorch base image (`nvcr.io\u002Fnvidia\u002Fpytorch:YY.MM-py3`) used by Megatron-LM CI. Covers the two pin sites (GitHub CI in `docker\u002F.ngc_version.dev` and GitLab CI in `.gitlab\u002Fstages\u002F01.build.yml`), the post-bump CI loop (re-run functional tests, refresh golden values, mark broken tests), and the gotchas that bit PRs",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8452,8455],{"name":8453,"slug":8454,"type":15},"CI\u002FCD","ci-cd",{"name":8439,"slug":8440,"type":15},"2026-07-14T05:25:59.97109",{"slug":8458,"name":8458,"fn":8459,"description":8460,"org":8461,"tags":8462,"stars":8443,"repoUrl":8444,"updatedAt":8468},"mcore-cicd","manage CI\u002FCD pipelines for Megatron-LM","CI\u002FCD reference for Megatron-LM. Covers CI pipeline structure, PR scope labels, triggering internal GitLab CI (which force-pushes the current branch to a pull-request\u002FBRANCH ref — always dry-run and verify the destination first; never run against shared or protected branches), and CI failure investigation.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8463,8464,8465],{"name":8453,"slug":8454,"type":15},{"name":8439,"slug":8440,"type":15},{"name":8466,"slug":8467,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":8470,"name":8470,"fn":8471,"description":8472,"org":8473,"tags":8474,"stars":8443,"repoUrl":8444,"updatedAt":8482},"mcore-create-issue","investigate CI failures and create issues","Investigate a failing GitHub Actions run or job and create a GitHub issue for the failure.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8475,8478,8479],{"name":8476,"slug":8477,"type":15},"Debugging","debugging",{"name":8466,"slug":8467,"type":15},{"name":8480,"slug":8481,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":8484,"name":8484,"fn":8485,"description":8486,"org":8487,"tags":8488,"stars":8443,"repoUrl":8444,"updatedAt":8495},"mcore-linting-and-formatting","lint and format Megatron-LM code","Linting and formatting for Megatron-LM. Covers running autoformat.sh, tools (ruff, black, isort, pylint, mypy), and code style rules.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8489,8492],{"name":8490,"slug":8491,"type":15},"Best Practices","best-practices",{"name":8493,"slug":8494,"type":15},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":8497,"name":8497,"fn":8498,"description":8499,"org":8500,"tags":8501,"stars":8443,"repoUrl":8444,"updatedAt":8507},"mcore-migrate-gpt-to-hybrid","migrate Megatron-LM models to HybridModel","Migration guide for moving Megatron Core GPTModel checkpoints, model providers, training commands, and layer mappings to HybridModel.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8502,8503,8506],{"name":23,"slug":24,"type":15},{"name":8504,"slug":8505,"type":15},"Migration","migration",{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":8509,"name":8509,"fn":8510,"description":8511,"org":8512,"tags":8513,"stars":8443,"repoUrl":8444,"updatedAt":8520},"mcore-onboard-gb200-1node-tests","onboard functional tests for GB200","Onboard 1-node GitHub MR functional tests for GB200 from existing mr-scoped 2-node tests.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8514,8517],{"name":8515,"slug":8516,"type":15},"QA","qa",{"name":8518,"slug":8519,"type":15},"Testing","testing","2026-07-14T05:25:53.673039",{"slug":8522,"name":8522,"fn":8523,"description":8524,"org":8525,"tags":8526,"stars":8443,"repoUrl":8444,"updatedAt":8531},"mcore-run-on-slurm","launch distributed training jobs on SLURM","How to launch distributed Megatron-LM training jobs on a SLURM cluster. Covers a minimal sbatch skeleton, environment-variable setup for torch.distributed.run, CUDA_DEVICE_MAX_CONNECTIONS rules across hardware and parallelism modes, container conventions, monitoring, and per-rank failure diagnosis.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8527,8528],{"name":8439,"slug":8440,"type":15},{"name":8529,"slug":8530,"type":15},"Infrastructure","infrastructure","2026-07-14T05:25:49.362534",{"slug":8533,"name":8533,"fn":8534,"description":8535,"org":8536,"tags":8537,"stars":8443,"repoUrl":8444,"updatedAt":8545},"mcore-split-pr","split pull requests to reduce review load","Split a PR into multiple PRs to reduce the number of required CODEOWNERS reviewer groups.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8538,8541,8542],{"name":8539,"slug":8540,"type":15},"Code Review","code-review",{"name":8466,"slug":8467,"type":15},{"name":8543,"slug":8544,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":8547,"name":8547,"fn":8548,"description":8549,"org":8550,"tags":8551,"stars":8443,"repoUrl":8444,"updatedAt":8554},"mcore-testing","run and manage Megatron-LM tests","Test system for Megatron-LM. Covers test layout, recipe YAML structure, adding and running unit and functional tests, golden values, marker filters, and CI parity.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8552,8553],{"name":8515,"slug":8516,"type":15},{"name":8518,"slug":8519,"type":15},"2026-07-14T05:25:54.928983",{"slug":8556,"name":8556,"fn":8557,"description":8558,"org":8559,"tags":8560,"stars":8443,"repoUrl":8444,"updatedAt":8565},"nightly-sync","manage nightly main-to-dev sync workflows","Domain knowledge for the nightly main-to-dev sync workflow. Covers merge strategy, CI architecture, failure investigation, and known issues.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[8561,8564],{"name":8562,"slug":8563,"type":15},"Automation","automation",{"name":8453,"slug":8454,"type":15},"2026-07-30T05:29:03.275638",496]