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

Datasetopenai/openai_humaneval

HumanEval: 164 Programming Tasks

A benchmark of 164 Python tasks with signatures, docstrings, and unit tests for code generation evaluation.

  • English
293.5K 420up to 1K rows
DatasetHuggingFaceM4/FineVisionMaxTop pick

Giant Collection of Images and Dialogues for VLMs

24.3 million interleaved image-text examples in English and Chinese for training multimodal models.

  • Image text → text
  • English
  • Chinese
291.9K 4810M–100M rows
Datasetgoogle-research-datasets/mbpp

MBPP Python Tasks Benchmark

About 1,000 Python tasks with descriptions, solutions, and tests for code generation evaluation.

  • English
288.7K 2791K–10K rows
DatasetRowan/hellaswag

HellaSwag Commonsense Benchmark

Multiple-choice sentence completion tasks to evaluate commonsense reasoning in language models.

  • English
282.2K 21710K–100K rows
Datasetlaion/voice-acting-cutscene-promptsTop pick

4M Cutscene Voice Acting Prompts

English and German text scripts with sudden emotional shifts for a single character to train expressive TTS models.

  • Text generation
  • English
  • German
273.7K 21M–10M rows
DatasetOmniAICreator/RoyalRoad-1.61M

1.61M Web Novels from Royal Road

Large-scale English corpus of 1.61 million web novel records containing text, tables, and images.

  • Text classification
  • Text generation
  • English
265.2K 61M–10M rows
DatasetXiaomiMiMo/MiMo-V2.6-RL-oss

RL Environments for LLM Agents

Environment suite for training LLM agents with checks in code, cybersecurity, web, and music.

    260.7K 1.1K1K–10K rows
    Datasetharborframework/terminal-bench-2.1

    Terminal-Bench 2.1 Benchmark for Harbor

    Mirror of Terminal-Bench 2.1 benchmark tasks adapted for execution in Harbor.

      260.4K 36up to 1K rows
      DatasetHubLabs/HubEgo-100KTop pick

      First-person real manual labor video dataset

      Over 10,000 hours of real-world recordings of craftsmanship, cooking, and construction work with IMU sensor data.

      • Robotics
      • Video classification
      • Video text → text
      • English
      257.2K 931B–10B rows🔒 on request
      Datasetespnet/yodas3Top pick

      YODAS v3 100+ language speech corpus

      Over 1.1 million hours of web audio in 100+ languages for ASR, TTS, and audio representation learning.

      • Audio → audio
      • Automatic speech recognition
      • Text → speech
      253.4K 2631M–10M rows
      Datasetdjghosh/wds_objectnet_test

      ObjectNet test split for object recognition

      A test split from the well-known ObjectNet benchmark for evaluating computer vision models without background and viewpoint bias.

        246.3K 1up to 1K rows
        Datasetmlfoundations/dclm-baseline-1.0Top pick

        DCLM-baseline: 4T tokens for pre-training

        4 trillion tokens and 3 billion documents for LLM pre-training; strong benchmark results.

          233.1K 3371B–10B rows
          Datasetespnet/yodas3.1Top pick

          Audio corpus in 100+ languages

          A massive dataset of over 1.1 million hours of audio across 100+ languages for speech tasks.

          • Audio → audio
          • Automatic speech recognition
          • Text → speech
          230.8K 2761M–10M rows
          Datasetgoogle/fleurs

          FLEURS multilingual ASR benchmark

          Speech data in 102 languages for training and evaluating ASR on parallel sentences.

          • Automatic speech recognition
          • Afrikaans
          • Amharic
          • Arabic
          • +9
          220.2K 541100K–1M rows
          DatasetHuggingFaceH4/MATH-500

          MATH-500 benchmark task subset

          A subset of 500 math problems from the MATH benchmark for evaluating step-by-step reasoning.

          • Text generation
          • English
          218.4K 358up to 1K rows
          Datasetroneneldan/TinyStories

          Short synthetic stories for LLM training

          Synthetic short stories with a limited English vocabulary for training and evaluating language models.

          • Text generation
          • English
          216.4K 1.3K1M–10M rows
          Datasetaoiandroid/android-times-articles

          Android Times multilingual article and transcript archive

          Synthesized and processed articles with metadata in 5 languages for text generation, translation, and summarization tasks.

          • Text generation
          • Summarization
          • Translation
          • English
          • Japanese
          • Chinese
          • +2
          213.9K 11K–10K rows
          Datasetieasybooks-org/waqfeya-library

          Digitized Islamic library in Arabic

          Over 10 thousand Arabic books in PDF, TXT, and DOCX formats, recognized via Google Document AI.

          • Image → text
          • Arabic
          210.6K 2310K–100K rows
          Datasetmlfoundations/MINT-1T-HTMLTop pick

          HTML portion of MINT-1T for multimodal pre-training

          English MINT-1T HTML corpus: part of an open 1 trillion token and 3.4 billion image multimodal dataset.

          • Image → text
          • Text generation
          • English
          209.3K 133100M–1B rows
          DatasetHuggingFaceFW/fineweb-2Top pick

          FineWeb-2 multilingual corpus

          FineWeb-2 pretraining corpus with 1000+ languages for training language models.

          • Text generation
          • aai
          • aak
          • aau
          • +9
          205.9K 9391B–10B rows
          Datasetmercor/apex-agentsTop pick

          APEX-Agents: AI agent benchmark

          480 tasks from bankers, consultants, and lawyers to test AI agents in cross-application scenarios.

          • English
          203.7K 225up to 1K rows🔒 on request