# VST (Visual Spatial Tuning) data

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

- id: `vst-visual-spatial-tuning-data`
- kind: language and reasoning annotation
- robot or device: single views, multiple images and videos (class: sensor only)
- how collected: constructed perception and reasoning QA
- size as stated: VST-P 4.1 million samples; VST-R 135K samples; Hub subset 500K
- hours: not stated | hours only claimed: none | episodes: not stated
- year: 2025
- organisation: rayruiyang (HF account)
- licence: not stated (class: not stated)
- access: open (class: open)
- link (dataset): https://huggingface.co/datasets/rayruiyang/vst_500k
- Hugging Face repository: `rayruiyang/vst_500k`, 4,333 downloads in 30 days, https://huggingface.co/datasets/rayruiyang/vst_500k
- note: Mirror allenai/Molmo2-ER-VST-P is tagged cc-by-nc-4.0.

Confirm the licence at the link before relying on it.