Real-world 3D research often lacks a practical bridge between recognizable retail objects and assets that can be loaded into interactive scenes. This collection makes Amazon Berkeley Objects meshes easier to use in browser-based experiments while retaining the original geometry in its highest-detail version. It is best understood as a compact, renderable asset layer for 3D perception work—not as a ready-made benchmark with task splits and labels.
AI Dataset2026
Amazon Berkeley Objects GLB
Provides 7,953 browser-optimized GLB models of real retail products, with PBR materials, product metadata, and multiple detail levels for 3D rendering and real-world object understanding research.
Introduction
Information
- Websitehuggingface.co
- OrganizationsCodePawl, Amazon.com, University of California, Berkeley
- AuthorsMatthieu Guillaumin, Thomas Dideriksen, Kenan Deng, Himanshu Arora, Arnab Dhua, Xi Zhang, Tomas Yago-Vicente, Jasmine Collins, Shubham Goel, Jitendra Malik
- Published date2026/11/04
Categories
More Items
AI Agent Papers2026
Turns a robot calibration episode and data request into synchronized, action-labelled training data: an agent builds the simulated task, while a video model renders realistic observations. Across five robot tasks, farm-trained policies scored 87 successes under unseen appearances, versus 46 for teleoperation-trained policies.
AI Agent Papers2026
Evaluates AI agents on 56 research-based physics tasks, from literature review and experimentation to manuscript writing. It combines deterministic checks with expert rubrics to assess both correctness and research quality.
AI Video Papers2026
Provides minute-scale continuous gameplay video, camera trajectories, and structured annotations for training interactive world models that remain coherent over longer rollouts. Its two-minute sequences address the shortage of long-horizon training data.