VSI-590K

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

• id: vsi-590k
• kind: language and reasoning annotation
• robot or device: image and video spatial QA from 3D ground truth and pseudo-labels (class: sensor only)
• how collected: curated and annotated from diverse sources
• size as stated: 590K samples
• hours: not stated | hours only claimed: none | episodes: not stated
• year: 2025
• organisation: nyu-visionx
• licence: apache-2.0 (class: open)
• access: open (class: open)
• link (dataset): https://huggingface.co/datasets/nyu-visionx/VSI-590K
• Hugging Face repository: nyu-visionx/VSI-590K, 4,146 downloads in 30 days, https://huggingface.co/datasets/nyu-visionx/VSI-590K
• note: Training set from the Cambrian-S paper.

Confirm the licence at the link before relying on it.