A 3D model can look complete yet still be unusable for robotics simulation: collision geometry, mass, friction, and other physical properties determine how an object behaves. This collection addresses that gap by making simulation metadata part of the asset, so teams can spend less time preparing scene contents and more time testing robot behaviors.
AI Dataset2026
SimReady Central Library
Provides more than 2,000 OpenUSD assets with properties needed for simulation, helping developers build robot-training environments and AI factory digital twins. Asset capabilities vary, so check profile compatibility for each workflow.
Introduction
Information
- Websitehuggingface.co
- OrganizationsNVIDIA
- Published date2026/09/29
Categories
More Items
Rewinds world-model trajectories to test alternative robot actions and turn outcome comparisons into action-level training signals. Real-robot data continually corrects the model; across three physical tasks, success rates beat the strongest baseline by 10.67–11.85 percentage points.
Conditions robot action generation on predicted visual and motion features, capturing physical dynamics without generating future video during control. Reports results across four simulation benchmarks.
AI Dataset2026
Provides self-contained synthetic environments for training data-analysis agents. This slice contains 1,538 tasks, each graded with a rubric designed to evaluate analytical findings.