GRAIL
A digital data-generation pipeline that synthesizes humanoid loco-manipulation data from 3D assets and video priors.
About
GRAIL is a fully digital data-generation pipeline for humanoid loco-manipulation. It composes 3D assets, simulator-ready scenes, robot-proportioned characters, and video foundation model priors to synthesize metric 4D human-object interaction (HOI) trajectories, then retargets them to a Unitree G1 and trains task-general policies for pick-up, whole-body manipulation, sitting, and terrain traversal. Using only GRAIL-generated data, the resulting egocentric visual policies transfer to real-world object pick-up and stair-climbing.
Open Source Health
- Stars
- 508
- Forks
- 39
- License
- Not stated
- Last commit
- 1 months ago
Resources & Links
Related Categories
Vendor
Nvlabs
Publisher of Pacnet, MUNIT and OmniVinci
Quick Links
Open Source
More by Nvlabs
Related Products
Matrix
Matrix (Multi-Agent daTa geneRation Infra and eXperimentation framework) is a versatile engine for multi-agent conversational data generation.
Top category match
DataArc-SynData-Toolkit
Synthetic Data Generation Platform By DataArcTech
Top category match
Pacnet
Pixel-Adaptive Convolutional Neural Networks (CVPR '19)
More from this vendor
MUNIT
Multimodal Unsupervised Image-to-Image Translation
More from this vendor
OmniVinci
OmniVinci is an omni-modal LLM for joint understanding of vision, audio, and language.
More from this vendor