Loading...
Loading...
3D models, rendered datasets, physics simulation, digital twins, synthetic data generation, game engine data
1,020 datasets
NVIDIA released the Physical AI NuRec dataset in March 2026 to facilitate end-to-end robotic model training. It provides 3D assets in Universal Scene Description (USD) format, including mesh components for collision detection and occupancy maps for navigation within the Isaac Sim environment.
9 MP4 video sequences of egocentric retail store footage captured by shopping cart cameras, created by Voxel51. Each frame includes 3-DOF camera pose (position and orientation) and store-zone classification labels for indoor localization tasks. The data includes synchronized RGB frames and depth maps from a real supermarket environment.
Harmony4D is a large-scale multi-view video dataset featuring in-the-wild close humanβhuman contact interactions. It includes activities like wrestling, dancing, MMA, karate, fencing, and hugging, with dense ground-truth annotations for detection, tracking, and 3D pose estimation. The dataset was created by Voxel51 and last updated on March 18, 2026.
A data bundle for machine learning workflows, including training, inference, and synthetic data components. The dataset's origin, scale, and creation date are unspecified. It was sourced from the Kaggle platform.
Irfanuruchi/building-engineering-synthetic-dataset-v5 is a synthetic dataset for training engineering reasoning models. It was generated using physics-based engineering equations and structured prompts suitable for LLM fine-tuning. The dataset was used to train the model Irfanuruchi/qwen2.5-1.5b-buildeng-precheck-lora-v5.
A three-dimensional musculoskeletal model of the Cambrian phosphatic brachiopod Genetreta trilix, used for biomechanical simulations. The Blender file includes reconstructed shell geometry, muscle meshes, attachment areas, and a kinematic setup. It was employed to simulate valve opening and estimate relative muscle forces for a related scientific publication.
A Blender file contains the complete three-dimensional musculoskeletal model of the Cambrian phosphatic brachiopod Iberotreta sampelayoi. This model includes reconstructed shell geometry, muscle meshes, attachment areas, and a kinematic setup used to simulate valve movement and estimate muscle forces. It supports the biomechanical analysis presented in the associated research on Cambrian brachiopod diversification.
Slighting3121's 3D Reconstruction Benchmark Asset Inspection dataset provides simulated scenes for evaluating 3D reconstruction and inspection algorithms. The dataset includes camera parameters, depth maps, and images under varying surface soiling conditions, as indicated by the file structure. It was last updated on March 10,ζ们εη°2026.
A 2023-2026 report from datadiscovery.nlm.nih.gov lists new Entry Combinations in the Medical Subject Headings (MeSH) vocabulary. The data, last updated on 2026-02-10, documents cases where new precoordinated Descriptors replace existing Descriptor and Qualifier pairs. It includes columns for DESCRIPTOR, QUALIFIER, ENTRY_COMBINATION_DESCRIPTOR, and their unique identifiers.
A synthetic dataset of 7,469 text pairs for training and evaluating models that predict dating compatibility. The dataset was created by dipenbhuva and last updated on March 9, 2026. It is designed for tasks like embedding fine-tuning and text pair classification.
A grouped dataset from FiftyOne containing the synthetic evaluation benchmark for Reflect3r, a 3DV 2026 paper on single-view 3D reconstruction. It includes 16 synthetic Blender interior scenes, each rendered from a real camera and a derived virtual mirror camera, accompanied by ground-truth point clouds. The dataset was created by Voxel51 and last updated on March 17, 2026.
A synthetic dataset designed for e-commerce risk analysis tasks. It likely contains features for classifying orders, such as fraud detection and return prediction. The dataset is hosted on Kaggle, but specific details about its creation, size, and update history are unknown.
Bin Picking - Bracket 02 is a synthetic dataset for robotics and computer vision. The dataset likely contains images or point clouds of mechanical brackets in cluttered industrial bins, simulating a common automation task. It was sourced from Kaggle, but details on its creator, size, and license are unknown.
A synthetic dataset depicting mechanical flanges in cluttered industrial bins. The dataset likely contains images or point clouds generated for computer vision tasks. It was sourced from Kaggle, but details on its creation date, author, and volume are unknown.
Synthetic data simulates mechanical brackets in cluttered industrial bins for robotic vision tasks. The dataset likely contains images or point clouds generated for training and testing bin picking algorithms. Its origin, size, and specific file formats are not detailed in the provided metadata.
A synthetic dataset for robotics and computer vision applications. The data likely contains images or point clouds of mechanical gears with 20 teeth arranged in cluttered industrial bins. The dataset's author, organization, size, and license are unknown.
Two datasets of surface-wave dispersion measurements for paths across the Pacific, used by Latallerie et al. (2024) to build a Vs model of the Pacific upper-mantle. Measurements are provided for source-receiver pairs across frequencies from 6 to 21 mHz, every 1 mHz, using a multi-taper technique with the first 5 Slepians.
A synthetic dataset for analyzing technology compensation. It likely contains information on salaries and associated skills within the tech industry. The dataset's author, organization, and specific size are unknown.
PointCloud_zips_ShapeNet_3D is a dataset of 3D point clouds, likely derived from the ShapeNet repository. It is hosted on Kaggle, but the exact size, number of objects, and creation details are unspecified. The dataset's content and structure require verification after download.
Synthetic Data V2 is a dataset hosted on HuggingFace by author margiela00. The dataset was last updated on April 21, 2026. Its specific content, size, and structure are not described in the available metadata.