{"success":true,"data":{"author":"Peter Sharpe, Carmelo Gonzales, Corey Adams, Mehdi Ataei, Sanjay Choudhry (NVIDIA)","bucketAccessType":"public","bucketName":"nvidia/AeroZoo","bucketPrefix":"","category":"Physics","cloudProvider":"huggingface","createdAt":"2026-11-16T02:38:29.510Z","description":"AeroZoo is a 3D aerodynamics dataset from NVIDIA (Sharpe, Gonzales, Adams, Ataei, Choudhry; arXiv:2611.17152) built to test whether neural surrogates solve boundary-value problems rather than memorising one shape family in one wind tunnel. It poses 3,816 distinct incompressible flow problems at Reynolds numbers from 1e3 to 1e8, solved with unsteady RANS using the Langtry-Menter transition SST model in OpenFOAM 14 and time-averaged over a fixed convective window. Bodies come from seven families (random blobs, Blender bluff and streamlined shapes, procedural cars, Zero-to-CAD parts, Airframe Foundry aircraft, streamlined bodies and wings) placed in eight enclosure configurations: rectangular boxes with open, ground-effect, confined, crossflow and dual-inlet layouts, opposed inlet/outlet spheres, symmetry hemispheres and 303 internal ducts, with no-slip or banded slip walls. Each sample stores the posed problem and time-mean solution as a PhysicsNeMo DomainMesh (memory-mapped TensorDict): about 23.8 billion interior query points in total (median 5.0M cells), surface pressure and wall shear stress, force coefficients, transition and stability labels (2,817 mean_stable), with 22 frozen train/validation/test splits for interior and boundary tasks. About 1.48 TB in tar shards, with index.json and SHA256SUMS. CC-BY-4.0, with 500 Zero-to-CAD parts under Apache-2.0; cite Sharpe et al. 2026.","displayTitle":"AeroZoo v1.2: 3,816 3D Incompressible RANS Flow Problems Across Seven Geometry Families (NVIDIA PhysicsNeMo)","id":"3a1bca00-651b-51aa-ab82-5b0b1dec0fee","license":"CC-BY-4.0","licenseEvidenceUrl":"https://huggingface.co/datasets/nvidia/AeroZoo","objectCountEstimate":3836,"probeStatus":"ok","referenceUrl":"https://arxiv.org/abs/2611.17152","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"nvidia/AeroZoo\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"nvidia/AeroZoo\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=nvidia, license:cc-by-4.0, size_categories:1K<n<10K, downloads=307"],"format":"mixed","generatedAt":"2026-11-16T02:38:35.571Z","layout":{"example":"zero-to-cad-provenance.csv","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"sample_path,case_id,body,part_uuid,source_dataset,source_revision,source_file,source_stl_sha256,source_center,coordinate_scale_to_meters,mirrored,rotation_quaternion_xyzw,translation_m,reference_length_m,pose_mode,surface_bands,remeshing,notice\r\nsamples/shard-000/case-0000000000000000026-3baf53c668de.tar,case-0000000000000000026-3baf53c668de,primary,8f135164-ee1c-a567-574c-8ffc00b9f9ef,ADSKAILab/Zero-To-CAD-100k,48bcf0a8c6fbfb47a27f5662007c24dff8d754ae,0391_8f135164.stl,68af5c69f707ad1652a07ffe4ead8704085f3","encoding":"utf8","key":"zero-to-cad-provenance.csv","truncatedAt":512}],"sampleObjects":["NOTICE","SHA256SUMS","assets/anatomy.png","assets/diversity-and-regimes.png","assets/ducts.png"],"sidecars":[{"bytes":"AeroZoo\nCopyright (c) 2026 NVIDIA Corporation\n\nGOVERNING DOWNLOAD TERMS: Use of the dataset is governed by the Creative\nCommons Attribution 4.0 International License (CC BY 4.0), and includes 500\nparts selected from Autodesk AI Lab's Zero-to-CAD 100k dataset which is\ngoverned by Apache License, Version 2.0.\n\n  CC BY 4.0:                    https://creativecommons.org/licenses/by/4.0/legalcode.en\n  Apache License, Version 2.0:  https://www.apache.org/licenses/LICENSE-2.0\n\nThe Zero-to-CAD parts belong to the geometry family zero-to-cad-diverse-500,\nfrom Hugging Face ADSKAILab/Zero-To-CAD-100k. The file NOTICE says what\nAeroZoo changed, which samples contain them, and holds the full Apache\nLicense 2.0 text. The full legal code of CC BY 4.0 follows.\n\n=======================================================================\n\nAttribution 4.0 International\n\n=======================================================================\n\nCreative Commons Corporation (\"Creative Commons\") is not a law firm and\ndoes not provide legal services or legal advice. Distribution of\nCreative Commons public licenses does not create a lawyer-client or\nother relationship. Creative Commons makes its licenses and related\ninformation available on an \"as-is\" basis. Creative Commons gives no\nwarranties regarding its licenses, any material licensed under their\nterms and conditions, or any related information. Creative Commons\ndisclaims all liability for damages resulting from their use to the\nfullest extent possible.\n\nUsing Creative Commons Public Licenses\n\nCreative Commons public licenses provide a standard set of terms and\nconditions that creators and other rights holders may use to share\noriginal works of authorship and other material subject to copyright\nand certain other rights specified in the public license below. The\nfollowing considerations are for informational purposes only, are not\nexhaustive, and do not form part of our licenses.\n\n     Considerations for licensors: Our public licenses are\n     intended for use by those authorized to give the public\n     permission to use material in ways otherwise restricted by\n     copyright and certain other rights. Our licenses are\n     irrevocable. Licensors should read and understand the terms\n     and conditions of the license they choose before applying it.\n     Licensors should also secure all rights necessary before\n     applying our licenses so that the public can reuse the\n     material as expected. Licensors should clearly mark any\n     material not subject to the license. This includes other CC-\n     licensed material, or material used under an exception or\n     limitation to copyright. More considerations for licensors:\n    wiki.creativecommons.org/Considerations_for_licensors\n\n     Considerations for the public: By using one of our public\n     licenses, a licensor grants the public permission to use the\n     licensed material under specified terms and conditions. If\n     the licensor's permission is not necessary for any reason--for\n     example, because of any applicable exception or limitation to\n     copyright--then that use is not regulated by the license. Our\n     licenses grant only permissions under copyright and certain\n     other rights that a licensor has authority to grant. Use of\n     the licensed material may still be restricted for other\n     reasons, including because others have copyright or other\n     rights in the material. A licensor may make special requests,\n     such as asking that all changes be marked or described.\n     Although not required by our licenses, you are encouraged to\n     respect those requests where reasonable. More considerations\n     for the public:\n    wiki.creativecommons.org/Considerations_for_licensees\n\n=======================================================================\n\nCreative Commons Attribution 4.0 International Public License\n\nBy exercising the Licensed Rights (defined below), You accept and agree\nto be bound by the terms and conditions of this Creative Commons\nAttribution 4.0 International Public License (\"Public License\"). To the\nextent this ","encoding":"utf8","key":"LICENSE","name":"LICENSE","size":19486,"truncated":true},{"bytes":"---\npretty_name: AeroZoo v1.2\nviewer: false\nlicense: cc-by-4.0\ntags:\n  - cfd\n  - computational-fluid-dynamics\n  - rans\n  - openfoam\n  - physicsnemo\n  - aerodynamics\n  - mesh\n  - point-cloud\n  - physics-ai\n  - arxiv:2611.17152\nsize_categories:\n  - 1K<n<10K\ntask_categories:\n  - other\n---\n\n# AeroZoo v1.2\n\n<p align=\"center\">\n  <img src=\"assets/zoo.png\" alt=\"Velocity slices across AeroZoo's seven geometry families and eight enclosure configurations\" width=\"100%\"/>\n</p>\n\nMost 3D aerodynamics datasets vary a single parametric shape family inside one\nfixed wind tunnel, allowing surrogate models to memorize coordinate-to-field\nshortcuts instead of solving boundary-value problems or generalizing out of\ndistribution. **AeroZoo** poses 3,816 distinct 3D incompressible flow problems\nacross Reynolds numbers from \\\\(10^3\\\\) to \\\\(10^8\\\\), solved with unsteady\nRANS (Langtry–Menter transition SST in OpenFOAM 14) and time-averaged over a\nfixed convective window.\n\n<p align=\"center\">\n  <img src=\"assets/posing.png\" alt=\"One car geometry placed inside four different box enclosures at the same reference speed and Reynolds number, producing four distinct velocity fields and force coefficients\" width=\"660\"/>\n</p>\n<p align=\"center\"><em>One geometry, four boundary-value problems: placing the same car at the same reference speed and Reynolds number into four enclosures shifts axial force from C<sub>x</sub> = 0.025–0.029 to 0.554 and vertical force from near zero to C<sub>z</sub> = +1.066.</em></p>\n\nEvery sample is stored in dimensional SI units as a PhysicsNeMo `DomainMesh`\n(a memory-mapped PyTorch `TensorDict` tree) that bundles the posed\nboundary-value problem with its time-mean solution across:\n\n- **Seven geometry families:** Blobs, Blender bluff, Procedural cars,\n  Zero-to-CAD, Airframe Foundry, Streamlined, and Blender streamlined (plus one\n  mascot sample outside every split).\n- **Eight enclosure configurations** across four enclosure families:\n  rectangular boxes with five boundary layouts (open, ground-effect, confined,\n  crossflow, and dual-inlet), opposed inlet/outlet spheres, symmetry\n  hemispheres, and internal ducts (with or without an internal body).\n- **Three body-wall layouts:** all no-slip, two contiguous slip/no-slip bands,\n  or three bands.\n\n**v1.2 is the official release**, exported 2026-10-08 from a population frozen\non 2026-10-02. Authors: Peter\nSharpe\\*, Carmelo Gonzales\\*, Corey Adams, Mehdi Ataei, and Sanjay Choudhry\n(NVIDIA; \\*lead authors).\n\n## Links\n\n- This dataset: <https://huggingface.co/datasets/nvidia/AeroZoo>\n- PhysicsNeMo (`DomainMesh`): <https://github.com/NVIDIA/physicsnemo>\n- The AeroZoo paper: <https://arxiv.org/abs/2611.17152>\n\n## Visual Overview\n\nEvery body stores time-averaged pressure on its remeshed boundary triangles\n(`no_slip` and `slip`) and wall shear stress on `no_slip`, spanning high-genus\norganic blobs, bluff shapes, procedural cars, CAD parts, aircraft, streamlined\nbodies, wings, and the mascot:\n\n<p align=\"center\">\n  <img src=\"assets/surfaces.png\" alt=\"3D surface renders of pressure coefficient and skin-friction coefficient across all seven geometry families and the mascot, showing slip bands in gray\" width=\"100%\"/>\n</p>\n\nAt matched sample counts, AeroZoo's seven geometry families span 25 to 390\neffectively distinct shapes out of 400 under the Vendi diversity score\n(compared with 13 for DrivAerML and 11 for SHIFT-SUV at 400 shapes, and 7 out\nof 126 for HiLiftAeroML). Across five decades of Reynolds number, boundary\nlayers transition from fully laminar near \\\\(Re \\approx 10^3\\\\) (tracking\nBlasius flat-plate friction) through transitional turbulent wedges near\n\\\\(Re \\approx 10^5\\\\) to fully turbulent at \\\\(Re \\ge 10^6\\\\) (tracking Prandtl\nturbulent friction):\n\n<p align=\"center\">\n  <img src=\"assets/diversity-and-regimes.png\" alt=\"Vendi shape diversity across AeroZoo geometry families compared with single-family datasets, alongside boundary-layer laminar fraction and skin-friction coefficient across five decades of Reynolds number\" width=\"100%\"/>\n</p>\n\n<p align=\"center\">\n  <img src=\"assets/transiti","encoding":"utf8","key":"README.md","name":"README.md","size":43531,"truncated":true}],"sourceObjectKey":"zero-to-cad-provenance.csv","starterCell":"import pandas as pd\ndf = pd.read_csv(\"hf://datasets/nvidia/AeroZoo/zero-to-cad-provenance.csv\")\ndf.head()","starterSamplePath":"assets/anatomy.png","totalObjects":3836},"summary":"AeroZoo is a 3D aerodynamics dataset from NVIDIA (Sharpe, Gonzales, Adams, Ataei, Choudhry; arXiv:2611.17152) built to test whether neural surrogates solve boundary-value problems…","title":"AeroZoo v1.2: 3,816 3D Incompressible RANS Flow Problems Across Seven Geometry Families (NVIDIA PhysicsNeMo)","updatedAt":"2026-11-16T15:53:36.782Z","visibility":"public","websiteUrl":"https://huggingface.co/datasets/nvidia/AeroZoo"}}