Grasp-Anything

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

• id: grasp-anything
• kind: grasping
• robot or device: generated images with planar grasp labels (class: sensor only)
• how collected: foundation model generation
• size as stated: 1 million samples; 3 million objects
• hours: not stated | hours only claimed: none | episodes: not stated
• year: 2023
• organisation: AIRV Lab (airvlab)
• licence: MIT (Hub tag) (class: open)
• access: gated (class: gated)
• link (dataset): https://huggingface.co/datasets/airvlab/Grasp-Anything
• Hugging Face repository: airvlab/Grasp-Anything, 866 downloads in 30 days, https://huggingface.co/datasets/airvlab/Grasp-Anything
• note: Download form and licence agreement required.

Confirm the licence at the link before relying on it.