Skip to content

Datasets

Hugging Face datasets with plain-language descriptions: what's inside and what models to train on them. 4,788 in the catalog so far, with new ones added every hour.

Data type

For which models

Find a dataset for your goal

Answered by Gemini 4 · the chat is kept only in this tab

Datasetwulin222/MME-Unify

MME-Unify benchmark for multimodal models

Evaluation set of images and English text across 10 tasks and 30 subtasks for testing U-MLLMs.

  • Multiple choice
  • Question answering
  • Visual question answering
  • English
32.7K 31K–10K rows
DatasetJingyuSun/counterfactual-vlm-benchmark-data

Counterfactual VLM benchmark data

Dataset for benchmarking counterfactual vision-language models (VLMs).

  • Visual question answering
31.7K 2
Datasetderek-thomas/ScienceQA

ScienceQA: Scientific questions with images

A multimodal English science question dataset with multiple-choice options and images.

  • Multiple choice
  • Question answering
  • Other
  • English
31.3K 24410K–100K rows
Datasethiyouga/geometry3k

Geometric image-text VQA pairs

A VQA dataset with geometric diagram images and English text question-answer pairs, containing about 1 to 10k examples.

  • Visual question answering
  • English
30.6K 871K–10K rows
DatasetAlbertoChestnut/telugu-ocr

Telugu book scans and OCR text pairs

About 25.5 thousand pairs of scanned book pages and human-verified Telugu text from Wikisource.

  • Image → text
  • Telugu
30.5K 1100K–1M rows
DatasetVillanovaAI/Multi-CoSyn-400K

Multilingual synthetic document VQA dataset

100K to 1M question-answer pairs with reasoning over text-dense images across 5 European languages.

  • Visual question answering
  • Italian
  • English
  • German
  • +2
30.4K 0100K–1M rows🔒 on request
DatasetKRAFTON/VLM-SubtleBenchTop pick

VLM fine-grained visual comparison benchmark

A collection of images and English questions to test VLM ability to find subtle differences in similar pictures.

  • Visual question answering
  • Image → text
  • English
29.9K 710K–100K rows
Datasetmultimodal-reasoning-lab/Zebra-CoT

Zebra-CoT visual reasoning chains

182k alternating text and image examples for step-by-step visual reasoning.

  • Multimodal
  • Image text → text
  • Visual question answering
29.9K 84100K–1M rows
Datasetapple/DataCompDR-1BTop pick

Apple synthetic captions and embeddings for DataComp-1B

Apple's extension of the DataComp-1B dataset with synthetic captions, embeddings, and metadata for image-text pairs.

  • Text → image
  • Image → text
  • English
29.4K 371B–10B rows
DatasetSilversorrow/SpineBench

Medical VQA benchmark for spine diseases

64,878 English QA pairs for 40,263 spine images to evaluate multimodal models.

  • Visual question answering
  • Multiple choice
  • English
29.2K 010K–100K rows
DatasetCASIA-IVA-Lab/VisReason

VisReason visual reasoning benchmark

A set of 1,505 visual questions to test multimodal model image reasoning capabilities.

  • Visual question answering
28.4K 01K–10K rows
Datasetalibaba-multimodal-industrial-ai/IndustryBench-MIPU

Industrial product feature extraction benchmark

Image and Chinese text dataset for evaluating MLLMs in multi-image attribute extraction.

  • Image → text
  • Visual question answering
  • Chinese
27.9K 810K–100K rows
DatasetPawChrona/PawChrona-MMQA

Pet Health and Behavior Multimodal Q&A Dataset

49K English and Chinese questions about cats and dogs with images, audio, video, and text.

  • Visual question answering
  • Question answering
  • English
  • Chinese
27.4K 11100K–1M rows🔒 on request
DatasetBoKelvin/SLAKE

SLAKE Medical Visual Question Answering Dataset

Medical image dataset with questions, answers, and semantic annotations in English and Chinese.

  • Visual question answering
  • English
  • Chinese
26.8K 6510K–100K rows
DatasetHuggingFaceM4/DocmatixTop pick

Documents and questions for Document VQA

Massive English dataset for Document VQA used in Idefics3 fine-tuning.

  • Visual question answering
  • English
26.8K 3191M–10M rows
DatasetLin-Chen/MMStar

Data-Leakage-Free Multimodal Benchmark

Curated English visual-text tasks for fair evaluation of large multimodal models.

  • Multiple choice
  • Question answering
  • Visual question answering
  • English
26.2K 551K–10K rows
Datasetmlfoundations/MINT-1T-PDF-CC-2023-50Top pick

MINT-1T: interleaved text and images from PDF and the web

Part of a massive 1 trillion token multimodal dataset for pre-training models in document and page understanding.

  • Image → text
  • Text generation
  • English
25.8K 141M–10M rows
DatasetDongping-Li/EMMOE-100

Robotics, RL, and VQA training dataset

An English dataset of 10K–100K examples featuring scene descriptions, trajectories, and VQA tasks for embodied AI.

  • Visual question answering
  • Reinforcement learning
  • Robotics
  • English
25.7K 110K–100K rows
DatasetThinkingHub/PPTop pick

SteelBench VLM industrial safety benchmark

A dataset of annotated CCTV clips for evaluating VLMs in industrial safety.

  • Video classification
  • Image classification
  • Visual question answering
  • English
25.4K 01K–10K rows
DatasetOneThink/OneThinker-train-dataTop pick

OneThinker-600k multimodal dataset

Large-scale multimodal data corpus for training a universal reasoning model.

  • Image text → text
  • Video text → text
  • Object detection
  • English
24.3K 1410K–100K rows
Dataset1fanj/Chartographer

Visual chart analysis benchmark

English chart-question pairs to evaluate whether VLMs actually analyze images instead of guessing.

  • Visual question answering
  • Question answering
  • English
23.9K 31K–10K rows
DatasetUCSC-VLAA/Recap-DataComp-1BTop pick

1B image-text pairs with re-captions

Large DataComp-1B image-text corpus with improved captions from LLaVA-1.5-LLaMA3-8B.

  • Zero shot classification
  • Text retrieval
  • Image → text
23.5K 2131B–10B rows
Datasetckchaos/ChartDiff

ChartDiff chart pair comparison benchmark

Large-scale benchmark for comparative summarization of chart pairs by language models.

  • Summarization
  • Image text → text
  • Image → text
23.1K 01K–10K rows
DatasetOmniEdu/Omni-Edu

Omni-Edu Multimodal Dataset

Instructional dataset for educational tasks and multimodal learning.

  • Text generation
  • Question answering
  • Image text → text
  • English
  • Chinese
23K 010K–100K rows