SDG-SynHuman (PhysicalAI-WorldModel-Synthetic-Digital-Human-Scenes)

From the catalogue. Listed by one researcher from one source page.

• id: sdg-synhuman-physicalai-worldmodel-synthetic-digital-human-scenes
• kind: video for world models
• robot or device: synthetic digital humans in 3D environments (class: simulation)
• how collected: rendered synthetic video
• size as stated: 236,937 clips, about 5,841 hours
• hours: 5,841 | hours only claimed: none | episodes: not stated
• year: 2026
• organisation: NVIDIA
• licence: other (class: custom terms)
• access: open (class: open)
• link (dataset): https://huggingface.co/datasets/nvidia/PhysicalAI-WorldModel-Synthetic-Digital-Human-Scenes
• Hugging Face repository: nvidia/PhysicalAI-WorldModel-Synthetic-Digital-Human-Scenes, 29,329 downloads in 30 days, https://huggingface.co/datasets/nvidia/PhysicalAI-WorldModel-Synthetic-Digital-Human-Scenes
• other names: PhysicalAI-WorldModel-Synthetic-Digital-Human-Scenes
• note: Built for Cosmos world model training.
• download links of one episode: https://datasets.gurasees.com/episodes/nvidia/PhysicalAI-WorldModel-Synthetic-Digital-Human-Scenes.md?episode=0

Confirm the licence at the link before relying on it.