BridgeEQA

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

• id: bridgeeqa
• kind: language and reasoning annotation
• robot or device: 200 real-world bridge inspection scenes (class: sensor only)
• how collected: QA grounded in professional inspection reports
• size as stated: 2,200 open-vocabulary QA pairs; 47.93 images per scene on average
• hours: not stated | hours only claimed: none | episodes: not stated
• year: 2026
• organisation: hoskerelab
• licence: mit (class: open)
• access: open (class: open)
• link (dataset): https://huggingface.co/datasets/hoskerelab/bridge-eqa
• Hugging Face repository: hoskerelab/bridge-eqa, 469 downloads in 30 days, https://huggingface.co/datasets/hoskerelab/bridge-eqa
• note: Civil bridges, unrelated to BridgeData robot dataset.

Confirm the licence at the link before relying on it.