datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
InsightVQA
InsightVQA: High-Dimensional Emotion-Cognitive Visual Question Answering Benchmark
Overview
InsightVQA is a large-scale dataset designed for hierarchical visual question answering that bridges emotion understanding and cognitive reasoning. While existing benchmarks predominantly focus on surface-level emotion recognition , InsightVQA introduces a structured paradigm to evaluate a model's ability to interpret emotional causes, ground evidence, and reason about… See the full description on the dataset page: https://huggingface.co/datasets/ziyul707/InsightVQA.cacherefinedweb-embed-english-v3.0SlideVQA
SlideVQA
SlideVQA: A Dataset for Document Visual Question Answering on Multiple Images
📖 arXiv 🌐 github
We introduce a new document VQA dataset, SlideVQA, for tasks wherein given a slide deck composed of multiple slide images and a corresponding question, a system selects a set of evidence images and answers the question.
Citation and contact
If you use this dataset, please cite our work:
@inproceedings{SlideVQA2023,
author = {Ryota Tanaka and… See the full description on the dataset page: https://huggingface.co/datasets/NTT-hil-insight/SlideVQA.empathic-insights-voiceInSight-doc-SFT-18k
InSight-doc-SFT-18k
Agentic Visual Perception for Long-Document Understanding
📄 Paper |
💻 Code |
🤗 Model |
🎯 RL Data |
🎬 Replay Demo |
🚀 Live Demo
Understand the big picture. Focus on the right details. Answer from the evidence.
InSight-doc-SFT-18k is the supervised fine-tuning corpus used to train the
InSight-doc long-document understanding agent. Each example is a complete
multimodal trajectory: the agent starts from low-resolution document… See the full description on the dataset page: https://huggingface.co/datasets/m-Just/InSight-doc-SFT-18k.InSight-doc-RL-19k
InSight-doc-RL-19k
Agentic Visual Perception for Long-Document Understanding
📄 Paper |
💻 Code |
🤗 Model |
🧩 SFT Data |
🎬 Replay Demo |
🚀 Live Demo
Understand the big picture. Focus on the right details. Answer from the evidence.
InSight-doc-RL-19k is the reinforcement-learning corpus used to
train the InSight-doc long-document understanding agent after SFT. It contains
challenging document VQA prompts, unanswerable negatives, multiple-choice… See the full description on the dataset page: https://huggingface.co/datasets/m-Just/InSight-doc-RL-19k.locomoinsight_bench
InsightBench: Evaluating Business Analytics Agents Through Multi-Step Insight Generation
Dataset Summary
[Paper][Website][Dataset]
Insight-Bench is a benchmark dataset designed to evaluate end-to-end data analytics by evaluating agents' ability to perform comprehensive data analysis across diverse use cases, featuring carefully curated insights, an evaluation mechanism based on LLaMA-3-Eval or G-EVAL, and a data analytics agent, AgentPoirot.
1. Install the… See the full description on the dataset page: https://huggingface.co/datasets/ServiceNow/insight_bench.InSight
InSight: A Benchmark for Agentic Claim Verification in Interactive Visualizations
InSight evaluates multimodal agents on claim verification over interactive visualizations. Given an interactive HTML visualization and a natural-language proposition, an agent must explore the visualization over multiple turns (clicking, hovering, navigating) and classify the proposition as True, False, or NotEnoughInfo.
Paper: arXiv:2609.01383
Code / benchmark harness:… See the full description on the dataset page: https://huggingface.co/datasets/maevehutch/InSight.INSIGHTposrand_colorInsightBench-Assetsrfi-detection-dataset
RFI Detection Dataset
Synthetic Aperture Radar (SAR) dataset for radio-frequency interference (RFI) detection, developed as part of the OpenSAR Insight project. See the organization card for full project background, funding, and consortium details.
Codebase: https://github.com/ESA-PhiLab/OpenSARInsight
Project page: https://opensarinsightweb.web.uah.es/rfi_detection.html
Companion model: opensar-insight/rfi-detection-model
Overview
This dataset uses SAR images… See the full description on the dataset page: https://huggingface.co/datasets/opensar-insight/rfi-detection-dataset.VDocRetriever-Pretrain-DocStructLight-INSIGHT-Bench
INSIGHT-Bench v1
A human-curated object-goal navigation benchmark: 1,097 episodes over 210 scenes, each
episode a short natural-language instruction, a start pose, a goal position and a success radius,
defined on Z-up, metre-scaled USD conversions of four scene sources -- HM3D, Matterport3D,
InteriorGS and Habitat-GS (3D Gaussian Splatting). It is evaluated in NVIDIA Isaac Sim by
the INSIGHT-Bench evaluation SDK, which publishes exactly one coordinate over these bytes:… See the full description on the dataset page: https://huggingface.co/datasets/LightOriginsHQ/Light-INSIGHT-Bench.LaSeRSEarthReason
EarthReason📂:
The first large-scale benchmark dataset for geospatial pixel reasoning
📥Dwonload Dataset:
git lfs install
git clone https://huggingface.co/datasets/earth-insights/EarthReason
📦Additional Resources:
paper: ArXiv
project: earth-insights/SegEarth-R1
🚀Citation:
@article{li2025segearth,
title={SegEarth-R1: Geospatial Pixel Reasoning via Large Language Model},
author={Li, Kaiyu and Xin, Zepeng and Pang, Li and Pang, Chao and… See the full description on the dataset page: https://huggingface.co/datasets/earth-insights/EarthReason.InsightEval
InsightEval
InsightEval is an expert-curated benchmark for assessing whether LLM-driven data agents can discover meaningful, evidence-grounded insights from tabular data.
Zhenghao Zhu*, Yuanfeng Song*, Xing Chen, Chengzhong Liu, Yakun Cui, Caleb Chen Cao, Sirui Han, and Yike Guo.InsightEval: An Expert-Curated Benchmark for Assessing Insight Discovery in LLM-Driven Data Agents. Findings of ACL 2026.
[Paper] [PDF] [Code]
Dataset summary
Existing insight-discovery… See the full description on the dataset page: https://huggingface.co/datasets/zhenghaozhu/InsightEval.OpenDocVQA
Dataset Card for OpenDocVQA
This is a training and evaluation QA data file for VDocRAG, a new RAG framework that can directly understand diverse real-world documents purely from visual features.
Dataset Description
OpenDocVQA is the first unified collection of open-domain document visual question answering datasets, encompassing diverse document types and formats.
Supported Tasks and Leaderboards
Given a large collection of document images and a question, the… See the full description on the dataset page: https://huggingface.co/datasets/NTT-hil-insight/OpenDocVQA.thinking-design-v3-harness
thinking_design_v3 — does any multi-turn harness beat single-pass ideation?
Arm outputs, harness turn traces, and judge verdicts for a matched-idea-budget comparison of
single-pass ideation against two multi-turn harnesses, on Qwen3-4B-SI and Qwen3-8B-SI over a
fixed 100-instance slice.
Layout
path
what
arms/arms_<model>_<arm>.jsonl
3-idea sets per instance. _goldlc = same candidates, gold compressed to the candidate length
turns/twoturn_h{2,4}_{4b… See the full description on the dataset page: https://huggingface.co/datasets/insightSynthesisData/thinking-design-v3-harness.INSIGHTposrandvessel-detection-dataset
Vessel Detection Dataset
Synthetic Aperture Radar (SAR) dataset for vessel detection, including dark/non-cooperative vessels not broadcasting AIS, developed as part of the OpenSAR Insight project. See the organization card for full project background, funding, and consortium details.
Codebase: https://github.com/ESA-PhiLab/OpenSARInsight
Project page: https://opensarinsightweb.web.uah.es/dark_vessel.html
Companion model: opensar-insight/vessel-detection-model… See the full description on the dataset page: https://huggingface.co/datasets/opensar-insight/vessel-detection-dataset.miraclveri-bilimci-insight-diyalog-tr-16.2k
🇹🇷 Veri Bilimci Insight Diyalog Veri Seti (TR, 16.2K) — %100 Türkçe Metin
Gerçek dünya blogları, uzman soru-cevap içerikleri ve akademik makale metinlerinden üretilmiş; veri madenciliği ve uygulamalı veri bilimi karar diline odaklanan, %100 Türkçe çok turlu diyalog veri seti.
🧠 Bu Veri Seti Ne Amaçla Üretildi?
Amaç, modeli teorik tanım ezberinden çıkarıp bağlama göre karar veren veri bilimci davranışına yaklaştırmaktır.
Her örnekte yöntem seçimi, alternatif kıyası… See the full description on the dataset page: https://huggingface.co/datasets/zero9tech/veri-bilimci-insight-diyalog-tr-16.2k.OVEarth-Bench
OVEarth-Bench
OVEarth-Bench is an evaluation benchmark for open-vocabulary Earth-observation image understanding. It evaluates whether a model can recognize, segment, and localize remote-sensing targets from category names, referring expressions, and reasoning-oriented queries. The benchmark additionally includes negative open-vocabulary queries to measure hallucination suppression.
The associated evaluation toolkit is available at earth-insights/OVEarth-bench.… See the full description on the dataset page: https://huggingface.co/datasets/earth-insights/OVEarth-Bench.insight-doc-rl-v0flood-detection-dataset
Flood Detection Dataset
Synthetic Aperture Radar (SAR) dataset for flood and water body detection, developed as part of the OpenSAR Insight project. See the organization card for full project background, funding, and consortium details.
Codebase: https://github.com/ESA-PhiLab/OpenSARInsight
Project page: https://opensarinsightweb.web.uah.es/flood_detection.html
Overview
This dataset uses SAR images from the ESA Copernicus Sentinel-1 mission (Sentinel-1A and… See the full description on the dataset page: https://huggingface.co/datasets/opensar-insight/flood-detection-dataset.dfs-glossary
DFS Glossary — Amharic & Afaan Oromoo
Expert-verified glossaries of Digital Financial Services (DFS) terminology in
Amharic (am) and Afaan Oromoo (om), published as structured,
machine-readable, openly-licensed data.
Open language infrastructure for two low-resource Ethiopian languages — for
developers, researchers, translators, and the financial-inclusion community.
Languages
Amharic (am, Ge'ez script) · Afaan Oromoo (om, Latin script)
Entries
87 Amharic + 86… See the full description on the dataset page: https://huggingface.co/datasets/shega-insight/dfs-glossary.Slide_Insight_Images
About this Dataset
This Dataset contains data from several Presentation Slides. For each Slide the following information is available:
key: recordID_pdfNumber_slideNumber
image: each presentation slide as an PIL image
Zenodo Records Information
This repository contains data from Zenodo records.
Records
Zenodo Record 10008464Authors: Moore, JoshLicense: cc-by-4.0
Zenodo Record 10008465Authors: Moore, JoshLicense: cc-by-4.0
Zenodo Record… See the full description on the dataset page: https://huggingface.co/datasets/ScaDSAI/Slide_Insight_Images.Slide_Insight_Images_v2
About this Dataset
This Dataset contains several Presentation Slides as part of the NFDI4BIOIMAGE project SlideInsight to gain insights into presentation slides through multimodal AI models.
For each Slide the following information is available:
key: recordID_pdfNumber_slideNumber
image: each presentation slide as an PIL image
The corresponding embeddings and metadata can be found in this Huggingface Dataset.
Zenodo Records Information
This repository contains… See the full description on the dataset page: https://huggingface.co/datasets/ScaDSAI/Slide_Insight_Images_v2.