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

DatasetARTPARK-IISc/Vaani-Atypical-Speech-Corpus

Vaani atypical speech corpus

Google Euphonia project data corpus to improve ASR systems for people with atypical speech.

  • Automatic speech recognition
  • Hindi
  • Marathi
  • Telugu
2.4K 41K–10K rows🔒 on request
Datasetbiglam/europeana_newspapers

Historical European newspapers

European newspaper texts from the 18th to early 20th century for training and research.

  • Text generation
  • German
  • French
  • Greek
  • +9
2.4K 6910M–100M rows
Datasetwltjr1007/DomainNet

DomainNet classification dataset

Large image classification dataset from the WILDS benchmark in English.

  • Image classification
  • Zero shot image classification
  • English
2.4K 7100K–1M rows
DatasetFreedomIntelligence/PubMedVision

Medical VQA based on PubMed articles

Large set of English medical image-question-answer pairs from PubMed.

  • Question answering
  • Text generation
  • English
2.4K 1071M–10M rows
Datasetfacebook/winoground

Winoground: compositional reasoning

Image and caption pairs with identical words in different orders to test vision-language models.

  • Image → text
  • Text → image
  • Image classification
  • English
2.3K 124up to 1K rows🔒 on request
Datasetwanderkid/UniMER_Dataset

Image-based mathematical formula recognition

Over a million images of complex mathematical expressions for training OCR and image-to-text systems.

  • Image → text
  • English
  • Chinese
2.3K 301M–10M rows
DatasetAI45Research/ATBench

AI agent trajectory safety benchmark

A set of long AI agent tool-use scenarios for evaluating and diagnosing agent safety.

    2.3K 461K–10K rows
    Datasetmoonshotai/WorldVQATop pick

    Visual World Knowledge Benchmark for MLLMs

    Image-based questions to evaluate atomic world knowledge in multimodal models.

    • Visual question answering
    • English
    • Chinese
    2.3K 671K–10K rows
    Datasetshunk031/STAIR-Captions

    Large Japanese image captioning dataset

    A large-scale STAIR Captions dataset with 820,310 Japanese image captions for text-to-image and image-to-text generation tasks.

    • Image → text
    • Japanese
    2.3K 7100K–1M rows
    Datasetcodepawl/smithsonian-3d-glb

    Smithsonian 3D Artifact Scans in GLB Format

    A collection of optimized 3D models of museum artifacts for 3D generation, gaming, and computer vision.

      2.3K 31K–10K rows
      DatasetMilaWang/SpatialEval

      Spatial intelligence benchmark for VLMs

      A benchmark for evaluating the spatial intelligence of large language and multimodal models across four areas.

        2.2K 610K–100K rows
        DatasetByteDance/MTVQA

        Multilingual Text-Image VQA (9 Languages)

        Image-text question-answer pairs in nine languages, including Russian, for evaluating and training multimodal models.

        • Visual question answering
        • Image → text
        • Multilingual
        • Arabic
        • German
        • +7
        2.2K 421K–10K rows
        Datasetmarmal88/skin_cancer

        Dermatoscopic Skin Lesion Images

        The HAM10000 collection of 10,000+ dermatoscopic images for skin lesion classification and segmentation.

        • Image classification
        • Image segmentation
        • English
        2.1K 4710K–100K rows
        Datasetafaji/cvqa

        Multilingual culturally-aware VQA benchmark

        Over 10,000 multilingual image-question pairs with cultural context for evaluating multimodal models.

        • Question answering
        • Indonesian
        • Sundanese
        • Japanese
        • +9
        2K 4010K–100K rows
        DatasetTIGER-Lab/ViRL39KTop pick

        Verifiable visual tasks for VLM RL training

        A collection of 38,870 cleaned and verified image-question-answer pairs for training reasoning multimodal models.

        • Question answering
        • Image text → text
        • English
        2K 43
        Datasetnvidia/Nemotron-Personas-IndiaTop pick

        Synthetic Indian personas with real demographics

        1 to 10 million synthetic Indian resident profiles in English and Hindi for data generation and role-playing models.

        • Text generation
        • English
        • Hindi
        2K 591M–10M rows
        Datasetazyg0s/pixelart-172k

        172k pixel art 32x32 icons with text descriptions

        A large dataset of game sprites on a transparent background with English captions for text-to-image training.

        • Text → image
        • English
        1.9K 24100K–1M rows