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

DatasetSalesforce/wikitext

WikiText: Wikipedia texts for language modeling

Corpus of verified English Wikipedia articles for training and evaluating language models.

  • Text generation
  • Fill mask
  • English
2.5M 8701M–10M rows
Datasetopenai/gsm8k

Grade school math word problems

8.5K diverse grade school math problems with step-by-step solutions for reasoning evaluation.

  • Text generation
  • English
1.3M 2.1K10K–100K rows
Datasetallenai/c4

C4 cleaned web corpus

Massive cleaned Common Crawl corpus with en, realnewslike, and multilingual mC4 variants.

  • Text generation
  • Fill mask
  • Afrikaans
  • Amharic
  • Arabic
  • +9
775.4K 70510B–100B rows
Datasetcais/mmlu

MMLU: Massive multitask benchmark

Multiple-choice QA benchmark across 57 subjects in English.

  • Question answering
  • English
612.7K 898100K–1M rows
Datasetallenai/ai2_arc

AI2 ARC: Science Questions

7,887 multiple-choice science questions for evaluating question-answering models.

  • Question answering
  • English
545.4K 4331K–10K rows
DatasetHuggingFaceFW/fineweb-eduTop pick

FineWeb-Edu educational web texts

1.3T tokens of filtered educational web pages in English for LLM training.

  • Text generation
  • English
459.9K 1.5K1B–10B rows
Datasetksolovev/fine-newsTop pick

Cleaned corpus of 635M news articles in six languages

Large-scale multilingual archive of news articles from 2016–2026 with quality filtering and deduplication.

  • Text generation
  • Text classification
  • Text retrieval
  • English
  • Spanish
  • Russian
  • +9
459.8K 53100M–1B rows
Datasetm-a-p/FineFineWebTop pick

FineFineWeb: domain-specific web corpus

Large English web corpus broken down by dozens of domains and iteration statistics for training and classification.

  • Text classification
  • Text generation
  • English
435.2K 2441B–10B rows
Datasetliwu/MNBVC

Massive Chinese MNBVC text corpus

Massive Chinese text corpus for training and evaluating language models.

  • Text generation
  • Fill mask
  • Chinese
383.7K 680
Datasetgoogle/physical-engineerings-last-examTop pick

Google engineering task benchmark

A set of complex multi-step tasks from licensed engineers for evaluating AI agents.

  • Text generation
  • Question answering
  • English
378.3K 191up to 1K rows🔒 on request
Datasetepfml/FineWeb-HQ

Curated FineWeb texts for LLM pretraining

Top 10% highest-quality and structured web texts from the FineWeb corpus for language model training.

  • Text generation
  • English
368.1K 361B–10B rows
DatasetHuggingFaceFW/finewebTop pick

Cleaned English web corpus

Over 18.5 trillion tokens of cleaned and deduplicated English web texts from CommonCrawl for LLM training.

  • Text generation
  • English
361.5K 3.8K10B–100B rows
DatasetTuxifan/UbuntuIRC

Ubuntu IRC channel logs

Uncleaned archive of correspondence from Ubuntu IRC chats for training text generation language models.

  • Text generation
351.2K 41M–10M rows
Datasetwikimedia/wikipedia

Wikipedia in all languages

Cleaned Wikipedia articles in multiple languages, with one train split per language.

  • Text generation
  • Fill mask
  • Abkhazian
  • Acehnese
  • Adyghe
  • +9
333.5K 1.8K10M–100M rows
Datasetallenai/MADLAD-400Top pick

Multilingual document corpus across 419 languages

A Common Crawl document dataset spanning 419 languages, audited and more heavily filtered than alternatives.

  • Text generation
327.1K 185over 1T rows
Datasetnureasoning/nuReasoningTop pick

Multimodal Reasoning Dataset for Autonomous Vehicles

Synchronized camera, LiDAR, and step-by-step reasoning data for training autonomous driving systems.

  • Visual question answering
  • Text generation
  • Robotics
311.7K 2910K–100K rows🔒 on request
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
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
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
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