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

Datasetstanfordnlp/web_questions

Web Fact Question Answering Dataset

6,642 English question-answer pairs focused on specific Freebase knowledge base entities.

  • Question answering
  • English
16.1K 451K–10K rows
Datasetopencsg/chinese-fineweb-edu-v2

Chinese Fineweb Edu V2 corpus

Large Chinese pre-training corpus; version V2 is marked as deprecated.

  • Text generation
  • Chinese
16K 75100M–1B rows
Datasetlmsys/mt_bench_human_judgments

MT-bench Human Judgment Dataset

3.3k expert pairwise preference annotations for six model responses to 80 MT-bench questions.

  • Question answering
  • English
15.8K 1501K–10K rows
DatasetMathLLMs/MathVision

MATH-Vision benchmark for multimodal mathematics

Visual math problems with images and text for evaluating VLMs.

  • Question answering
  • Multiple choice
  • Visual question answering
  • English
15.7K 1841K–10K rows
Datasetopenbmb/UltraData-SFT-2605

MiniCPM5 SFT Data: Math and Code

Over 15M English and Chinese SFT examples for post-training the MiniCPM5-1B-SFT model.

  • Text generation
  • Question answering
  • English
  • Chinese
15.6K 42010M–100M rows🔒 on request
Datasettasksource/bigbench

BIG-Bench without heavy dependencies

Repackaged BIG-Bench for loading without TensorFlow and unnecessary dependencies.

  • Multiple choice
  • Question answering
  • Text classification
  • English
15.5K 70100K–1M rows
DatasetMAmmoTH-VL/MAmmoTH-VL-Instruct-12M

12M visual instructions for multimodal models

Image-instruction pairs for training visual assistants in English.

  • Visual question answering
  • Question answering
  • English
15.4K 6710M–100M rows
Datasetopenai/MMMLU

MMMLU: MMLU in 14 languages

MMLU test questions across 57 topics translated into 14 languages by professional translators.

  • Question answering
  • Arabic
  • Bangla
  • German
  • +9
15.4K 539100K–1M rows
Datasetmuyleanging/Akkhara-Sing-Khmer

Sing Khmer and Khmer script text pairs

A dataset of Khmer phrase pairs written in Latin script (Sing Khmer) and original script.

  • Text generation
  • Translation
  • Khmer
  • English
15.2K 3100K–1M rows🔒 on request
DatasetPKU-Alignment/PKU-SafeRLHF

Preferences and Q-A for LLM safety

Data for training and evaluating safe responses: preferences, question-answer pairs, and English prompts.

  • Text generation
  • English
14.9K 207100K–1M rows
DatasetSkywork/SkyPile-150B

Large Chinese web corpus for LLMs

Large filtered and deduplicated Chinese web text dataset for language model pre-training.

  • Text generation
  • Chinese
14.8K 4211M–10M rows
Datasetucinlp/drop

DROP: Discrete reading benchmark

96k paragraph-based questions to test arithmetic, counting, and sorting.

  • Question answering
  • English
14.6K 7210K–100K rows
DatasetCohereLabs/include-base-44

Multilingual benchmark of examination tests

22,637 four-choice questions across 44 languages to evaluate LLM knowledge and logic.

  • Multiple choice
  • Albanian
  • Arabic
  • Armenian
  • +9
14.4K 5310K–100K rows
Datasetgoogle/svq

Short Spoken Questions in 17 Languages

Audio questions across 26 locales to evaluate ASR, audio search, and audio embeddings.

  • Question answering
  • Automatic speech recognition
  • Text retrieval
  • Arabic
  • Bangla
  • English
  • +9
14.1K 681M–10M rows
Datasetbigcode/the-stack-dedup

Deduplicated source code corpus

Code in many programming languages with permissive licenses after near-deduplication, about 1.5 TB.

  • Text generation
  • Code
14.1K 431100M–1B rows🔒 on request
Datasetallenai/tulu-3-sft-personas-instruction-following

SFT instructions with personas and constraints

29,980 synthetic examples for training precise adherence to instructions with constraints from IFEval.

  • Text generation
  • English
14.1K 7110K–100K rows
DatasetShareGPT4Video/ShareGPT4Video

ShareGPT4Video: video captions

4.8M multimodal GPT-4-Vision video descriptions for LVLM and text-to-video models.

  • Visual question answering
  • Question answering
  • English
14K 21610K–100K rows
Datasetnvidia/Nemotron-Pretraining-Specialized-v1Top pick

Specialized pretraining data for Nemotron

Updated corpus for Nemotron 3: math, code, English web, and synthetic data.

  • Text generation
13.9K 9110M–100M rows
Datasetnvidia/Nemotron-CC-Math-v1Top pick

Math corpus from Common Crawl

133 billion math and code tokens from Common Crawl for LLM pre-training.

  • Text generation
13.8K 105100M–1B rows🔒 on request
DatasetHuggingFaceGECLM/REDDIT_comments

Comments from 50 popular Reddit subreddits

A massive corpus of English Reddit comments (2006-2023) for training language and dialogue models.

  • Text generation
  • English
13.8K 48100M–1B rows
DatasetHuggingFaceH4/stack-exchange-preferences

Stack Exchange preferences for preference tuning

Stack Overflow questions and answers with at least two responses for preference models and instruction tuning.

  • Question answering
  • English
13.7K 14010M–100M rows