NVIDIA Omniverse Replicator

NVIDIA Omniverse™ Replicator is a core extension of the Omniverse platform that lets developers easily build custom synthetic data generation (SDG) tools and pipelines. You now have a faster way to generate physically accurate synthetic data for training computer vision AI models and networks.

Download Omniverse Get Started Building

NVIDIA Omniverse Replicator SDK — Omniverse Replicator is available within Omniverse Code and is built on open standards such as Universal Scene Description (OpenUSD), Material Definition Language (MDL), and NVIDIA PhysX™.

Benefits of Developing with Omniverse Replicator

Accelerated AI Training

Bootstrap your computer vision model training process by quickly prototyping with synthetic data across a multitude of domains before training on real data.

Accurate AI Models

Increase model accuracy by generating diverse, photoreal, and physically accurate training data by varying lighting, object poses, scenes, and more.

Scale Up and Out

Generate large datasets quickly by easily scaling across multi-GPU, multi-node compute resources—either on premises or in the cloud.

Privacy and Diversity

Address privacy issues and reduce bias by generating diverse datasets to represent the real world.

Building Synthetic Data Pipelines

Multiple domains use Omniverse Replicator to build robust, iterative synthetic data generation pipelines.

Defect Detection

Robots

Self-Driving Cars

Custom Applications

Developer Features

Import Assets From Your 3D Content Tools

Speed up your data-generation pipeline by leveraging Omniverse Connectors from existing Universal Scene Description (OpenUSD)-compatible 3D Digital Content Creation (DCC) tools.

Generative AI Tools in Omniverse Replicator

Low-Code Workflow Unlocks Faster Content Generation

YAML-based workflow abstracts away the complexity of building your SDG pipelines. Describe the parameters and requirements of your scene using simple syntax to generate the data without writing extensive lines of code.

Watch the Replicator YAML Video

Create Diverse Data With Domain Randomization

Customizable, open APIs for randomizers and annotators let you build your own domain-specific workflow to programmatically generate diverse training data sets.

Read the Blog About Domain Randomization Watch the Tutorial on How to Randomize a Scene Using Replicator

Custom and Non-Visual Sensors

Easily simulate custom sensors at scale, simultaneously in both the visible and non-visible spectrum in the same scene (e.g. infrared).

Read the Blog About Physically-Based Sensor Simulations

Integrate Replicator Into Your Cloud-Native Workflows

Build custom synthetic data pipelines on a modern cloud-native infrastructure.

Self-Deploy and Manage Replicator in the Cloud Learn More About Omniverse Cloud Enterprise Offering

Defect detection for industrial digitalization

Generate Synthetic Data for Defect Detection

Learn how companies are using Omniverse Replicator to generate 3D synthetic data to detect defects in industrial digitalization.

Read the Blog Watch the Tutorial

Inspect Your Data With Replicator Insight

Available in the Omniverse Launcher, the Omniverse Replicator Insight app enables developers to quickly view, navigate, and inspect their synthetically generated renders.

Apply for Early Access

Built on Omniverse Replicator

NVIDIA and the developer ecosystem have built domain-specific synthetic data generation tools on Omniverse Replicator. Explore ready-made tools and get started on your synthetic data workflows today.

Self-Driving Cars

DRIVE Sim

NVIDIA DRIVE Sim™ is a high fidelity, physically based platform for end-to-end autonomous vehicle simulation. DRIVE Sim leverages the capabilities of Omniverse Replicator to generate pixel-accurate ground truth synthetic data for training, testing, and validating autonomous vehicle perception algorithms.

Explore DRIVE Sim

Robotics

Isaac Sim

NVIDIA Isaac Sim™ is a scalable robotics simulation application and synthetic data-generation tool that powers photorealistic, physically accurate virtual environments to develop, test, and manage AI-based robots. Isaac Sim uses Omniverse Replicator as its powerful synthetic data-generation engine.

Explore Isaac Sim

Resources

Get started building a synthetic data generation pipeline.

Blog

Learn how Omniverse Replicator can scale training and performance of AI perception networks.

Read the Blog

Training

Learn how to generate synthetic data for training computer vision models with Omniverse Replicator.

Take the Self-Paced Course

Documentation

Consult the Omniverse Replicator documentation.

Consult the Documentation

Tutorials

Dive into the step-by-step Omniverse Replicator tutorials.

Watch the Tutorials

Access

Deployment Options

On Premises

NVIDIA RTX Systems

The unrivaled power and capability found in the latest generations of NVIDIA RTX™-enabled GPUs deliver the performance needed for the real-time, multi-app, and AI features built into Omniverse.

Learn More About NVIDIA RTX Systems

Cloud

Omniverse Cloud

NVIDIA Omniverse™ Cloud is a cloud services platform that enables development, deployment, and management of advanced 3D applications and pipelines. Omniverse Cloud developer APIs are coming soon.

Learn More About Omniverse Cloud