Replicator#
Isaac Sim Replicator provides extensions and workflows for synthetic data generation (SDG) using the omni.replicator extension. It supports domain randomizations, sensor simulation, and data collection with annotators and writers. In contrast to omni.replicator, where the typical workflow combines randomization and data capture within a generated OmniGraph workflow, the Isaac Sim Replicator workflow focuses on flexibility by independently triggering randomizations and data captures. This makes it better-suited for nondeterministic scenarios such as simulations and robotics tasks.
Basics and Getting Started#
SDG Tutorials#
- Scene Based Synthetic Dataset Generation
- Object Based Synthetic Dataset Generation
- Environment Based Synthetic Dataset Generation with Infinigen
- Online Generation
- Pose Estimation Synthetic Data Generation
- Training Pose Estimation Model with Synthetic Data
- Randomization in Simulation – AMR Navigation
- Randomization in Simulation – UR10 Palletizing
- Custom Replicator Randomization Nodes