Sign inGet started
Dataset · Hugging Face

ONE-Lab/LiveVQA-new

LiveVQA 🏠 Project Page | 💻 GitHub | 📖 Paper LIVEVQA is a new dataset for visual knowledge seeking and updating, comprising 12 different categories, 28,488 visual entity and 107,138 cross-modality multi-hop reasoning questions. 🧠 Overview LIVEVQA is a benchmark dataset of over 107,000 samples designed to evaluate Multimodal Large Language Models (MLLMs) on up-to-date real-world visual information. Each sample pairs a recent image—collected from news, YouTube… See the full description on the dataset page: https://huggingface.co/datasets/ONE-Lab/LiveVQA-new.

Author / organizationONE-Lab
License (as declared)cc-by-nc-4.0
Downloads400 (as reported by Hugging Face)
Likes1
Created16 June 2025
Last updated at source7 July 2025
multimodalvqaknowledge-updatingretrieval-augmented-generationpeft
Embed this entry
<iframe src="https://silytics.si/embed/item/d516d82f-cd2e-40f5-ade6-8cc850618599" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Free, no tracking. More about feeds and embeds.

Recent changes

  • 16 Oct 2026 Description: LiveVQA 🏠 Project Page | 💻 GitHub | 📖 Paper LIVEVQA is a new dataset for visual knowledge seeking and updating, comprising 12 different categories, 28,488 visual entity and 107,138 cross-modality multi-hop reasoning questions. 🧠 Overvie → LiveVQA 🏠 Project Page | 💻 GitHub | 📖 Paper LIVEVQA is a new dataset for visual knowledge seeking and updating, comprising 12 different categories, 28,488 visual entity and 107,138 cross-modality multi-hop reasoning questions. 🧠 Overvie

Changes detected by Silytics between refreshes of the source.

More from ONE-Lab

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 16 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.