NVIDIA/GR00T-Dreams
Generates synthetic robot training data from single images and text
View on index · View in 3D Map
Related repos
- NVIDIA Isaac GR00T lets robots use language and images to do tasks in different environments
- Interactive world model from large-scale human videos
- Unified framework for robotics data collection
- Generates synthetic datasets using LLMs and statistical samplers
- Converts SOMA motion to robot animation
- Train and deploy models for agents
- CUDA-accelerated visual SLAM for robotics
- Zero-shot world action models for robots
- Open-source robotic arm kit with NVIDIA Jetson integration
- text-to-image model for fast high-res AI image generation
