DataSalon

Discover quality datasets for AI training — aggregated from 40+ platforms, curated by AI.

ProductSearch Datasets Browse Topics Rankings Community API / MCP

ResourcesDocumentation Blog Changelog Status

LegalPrivacy Policy Terms of Service Cookie Policy

All Categories

🔗

Multimodal & LLM

Image-text pairs, instruction tuning, visual QA, cross-modal data, foundation model training data

1,540 datasets

Multimodal & LLM

Multimodal Video Annotation Samples with Abstracted Visual Assets

SuperviseLab provides professional video annotation data for training multimodal AI models. This public sample dataset demonstrates annotation methodology and output quality across diverse video content categories. All visual assets have been abstracted to protect source privacy, and identifiable metadata has been removed.

VideoMultimodalVideo AnnotationTraining DataMultimodal AiComputer Vision+1

0 views

Multimodal & LLM

Spectra: Multimodal VQA Training Data for Science and Open-World Knowledge

Spectra is a multimodal question-answering training dataset designed for vision-language models. It combines graduate-level science questions from TQA and ScienceQA with open-world knowledge questions from OKVQA and science questions across physics, chemistry, math, and biology from AI2D. The dataset was created by Tamalmajumder and was last updated on April 18, 2026.

MultimodalTraining DataOpen World KnowledgeMultimodal VqaScience Questions+1

0 views

Multimodal & LLM

CVQAD: 1,962 FullHD Videos for Compression Artifact Evaluation

1,962 FullHD videos with YUV420 encoding and durations of 10-15 seconds form the open part of the MSU compression artifacts dataset. The dataset, developed by deepfakesMSU, includes videos at frame rates of 24, 25, 30, 39, 50, and 60 fps for evaluating video quality metrics. The full description is available on the dataset page, and the dataset was last updated on April 14, 2026.

VideoNo Reference MetricsBenchmarkComputer VisionFull Reference MetricsVideo QualityCompression Artifacts+1

0 views

Multimodal & LLM

DanQing100M: 100 Million Chinese Image-Text Pairs for Vision-Language Pre-training

DanQing100M is a large-scale Chinese vision-language dataset containing 100 million image-text pairs, totaling 12 terabytes. It was created by researchers including Hengyu Shen, Tiancheng Gu, and others from DeepGlint-AI, using web data from 2024 to 2025. The dataset is intended for vision-language pre-training tasks.

MultimodalChineseImage Text PairsWeb DataPre TrainingVision LanguageComputer VisionLarge Scale+1

0 views

Multimodal & LLM

GTPBD-MM: Multimodal Remote Sensing Data for Terraced Parcel Extraction

GTPBD-MM is the first multimodal benchmark for terraced scenes, integrating optical imagery, textual descriptions, and Digital Elevation Model (DEM) data. The dataset provides three levels of annotations: parcel, mask, and boundary. It was created by author wxqzzw and last updated on April 15, 2026.

GeospatialMultimodalLand ParcelsBenchmarkTerraced Agriculture+1

0 views

Multimodal & LLM

Multimodal Face Generation Data With Spatial And Semantic Conditioning

MMFace-DiT Dataset provides multimodal conditioning data for high-fidelity, controllable face synthesis. The dataset, created by BharathK333, includes spatial elements like masks and sketches paired with VLM-enriched semantic captions. It was accepted to CVPR 2026 and last updated in April 2026.

MultimodalComputer VisionFace GenerationSynthetic DataMultimodal Benchmark+1

0 views

Multimodal & LLM

FashionMV: Multi-View Fashion Images for Composed Image Retrieval

FashionMV is a large-scale dataset for product-level Composed Image Retrieval (CIR) created by yuandaxia. It contains 127,000 products, 472,000 multi-view images, and over 220,000 CIR triplets, built through an automated pipeline leveraging large multimodal models. The dataset was last updated on April 14, 2026.

MultimodalFashionComputer VisionLarge Scale+1

0 views

Multimodal & LLM

Hateful Memes Fine-Grained Dataset with 2,030 Multimodal Examples

2,030 memes form a fine-grained extension of the Hateful Memes dataset, annotated for nuanced analysis of harmful content. It was created by nils-herrmann and last updated on 2026-04 08. The dataset introduces annotation dimensions for incivility and intolerance beyond binary hatefulness.

MultimodalHate SpeechContent ModerationMeme Analysis+1

0 views

Multimodal & LLM

SightAct-Bench: A 14-Family Synthetic Benchmark for VLM Agent Safety

SightAct-Bench is a synthetic benchmark containing 14 families of tasks for evaluating the safety of Vision-Language Model-powered browser agents. The benchmark, authored by SightAct, Bench, and hosted on Harvard Dataverse, was last updated on 2026-05-04. It specifically tests whether agents safely handle task-relevant sensitive-information requests when a visually suspicious interaction is embedded in the workflow.

MultimodalVision Language ModelsBenchmarkSafety EvaluationBrowser AgentsSynthetic BenchmarkSynthetic+1

0 views

Multimodal & LLM

Computer Use Large: Professional Software Screen Recordings

Computer Use Large is a massive dataset containing 48,478 screen recording videos totaling approximately 12,300 hours of professional software usage. Sourced from the internet, the dataset is designed to facilitate research into GUI interaction and computer-use automation. It covers a variety of complex software environments including CAD tools, spreadsheets, and development environments.

Size Categories10 Kn100 KComputer UseLanguageenTask CategoriesroboticsScreen RecordingLicensecc By 40DesktopRegionusSoftware TutorialsTask Categoriesvideo ClassificationGui+1

0 views

Multimodal & LLM

Metis-RL: Reinforcement Learning Dataset for Multimodal Agent Training

Metis-RL contains approximately 5.2 thousand multimodal prompts designed for training agentic models. The dataset was created by Accio-Lab for reinforcement learning via Hierarchical Decoupled Policy Optimization and was last updated in April 2026. It focuses on cultivating meta-cognitive tool use across perception, search, and mathematical or logical reasoning tasks.

MultimodalTool UseMultimodal AiMeta CognitionReasoningReinforcement Learning+1

0 views

Multimodal & LLM

BEAT: Visual Backdoor Attack Training Data for Embodied AI Agents

BEAT is a training dataset for backdoor attacks on vision-language model-based embodied agents, accompanying a 2026 arXiv paper. The dataset is gated and intended for academic research on AI security and robustness. It was created by the UIUC Kang Lab and last updated on April 20, 2026.

MultimodalComputer VisionVisual language modelAi Security+1

0 views

Multimodal & LLM

BEAT: Visual Backdoor Attack Training Data for Embodied AI Agents

BEAT is a training dataset accompanying the research paper 'BEAT: Visual Backdoor Attacks on VLM-based Embodied Agents via Contrastive Trigger Learning'. The dataset is authored by Qiusi Zhan and collaborators and was last updated on April 20, 2026. It is hosted on Hugging Face and is gated, requiring access requests for academic research on AI security.

MultimodalContrastive LearningVisual Language ModelsSecurity Research+1

0 views

Multimodal & LLM

Instruction-Based Image Editing Dataset for Unified Multimodal Models

ScaleEdit-12M is a large-scale dataset for instruction-based image editing, created by InternVL-U to address challenges in constructing diverse, high-quality data without proprietary APIs. It was last updated on March 24, 2026. The dataset aims to overcome limitations of previous datasets that relied on closed-source models or fixed synthetic pipelines.

MultimodalParquetLibrarypolarsLibrarydaskLanguageenSize Categories10 Mn100 MModalitytextInternvlModalitytabularLibrarymlcroissantLibrarydatasetsComputer VisionInstruction Based EditingRegionusTask Categoriesimage To ImageArxiv260309877ScaleeditArxiv260320644Licensemit+1

0 views

Multimodal & LLM

GMAI-VL-5.5M: 5.5 Million Medical Vision-Language Question-Answering Pairs

Over 5.5 million multimodal question-answering pairs for training medical AI foundation models. The dataset was constructed by General-Medical-AI from hundreds of medical classification, segmentation, and detection datasets. It was last updated on the platform in April 2026.

MultimodalMultimodal QaFoundation ModelsMedical Vision LanguageHealthcareComputer VisionLarge ScaleMedical Ai+1

0 views

Multimodal & LLM

UMM: Unified Multimodal Model Results for the UEval Benchmark

UMM results for the UEval benchmark, a collection of outputs from various multimodal and large language models. The dataset was created by author wenwenw945 and last updated on April 9, 2026. It includes results from models like OmniGen2 and Emu3.5, configured for specific understanding tasks.

MultimodalModel EvaluationMultimodal LlmBenchmarkUnderstanding TasksBenchmark ResultsComputer Vision+1

0 views

Multimodal & LLM

Gemma Cvqa V4 Pred: Computer Vision Question Answering Predictions

Gemma Cvqa V4 Pred is a dataset uploaded by pltops to Hugging Face. The dataset's title suggests it contains predictions for computer vision question answering tasks, likely generated by or for the Gemma language model. It was last updated on 2026-05-23.

MultimodalQuestion AnsweringLlm EvaluationComputer Vision+1

0 views

Multimodal & LLM

Computational Target Analysis for Chronic Myelogenous Leukemia Therapy

Supplementary Material 2 from a study on multimodal gene and targeted drug therapy for chronic myelogenous leukemia. The 11.8 KB XLSX file contains supporting data for computational target analysis and therapeutic validation. Author Margaret L. Lugin published the file on figshare in April 2026.

TabularExcelDrug Target AnalysisNanoparticle TherapyComputational BiologyLeukemia Research+1

0 views

Multimodal & LLM

FORGE: Fine-grained Multimodal Evaluation for Manufacturing Scenarios

FORGE is a dataset for fine-grained multimodal evaluation in manufacturing scenarios, created by AI4Manufacturing. It includes at least 451 cases for tasks like wrong model detection, with data modalities including photos. The dataset was last updated on Hugging Face in April 2026.

MultimodalBenchmarkComputer VisionQuality controlMultimodal EvaluationManufacturing+1

0 views

Multimodal & LLM

CiQi-VQA: Visual Question Answering on Chinese Porcelain Artifacts

A multimodal dataset for cultural reasoning on antique Chinese porcelains, created by SII-Monument-Valley. The dataset is part of the CiQi-Agent project, which aligns visual perception, tool-augmented reasoning, and cultural knowledge. It was last updated on April 1, 2026.

MultimodalSize Categories10 Kn100 KTask Categoriesimage Text To TextLanguagezhTask Categoriesquestion AnsweringArxiv260328474LanguageenModalityimageBenchmarkComputer VisionLicensecc By Nc 40RegionusChinese ArtArtAgentMultimodal VqaCultural HeritageDomain AgentVisual Reasoning+1

0 views

PreviousPage 15 of 77Next

Multimodal & LLM Datasets | DataSalon