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

DatasetTeichAI/DeepSeek-v4-Pro-Agent

DeepSeek v4 Pro Agent Traces

4,006 raw agent traces with DeepSeek v4 Pro responses and a tool schema snapshot for SFT.

  • Text generation
  • English
9.6K 1111K–10K rows
Datasetnvidia/ToolScale

Synthetic tool-calling tasks

Tasks for RL training of small tool-orchestrator models and agents.

  • Text generation
  • English
9.6K 2101K–10K rows
Datasetfacebook/natural_reasoning

NaturalReasoning: reasoning tasks

A large set of complex questions with reference answers and generated solutions.

  • Text generation
  • English
9.6K 6001M–10M rows
Datasetgoogle/simpleqa-verified

Google short factual questions benchmark

1,000 verified English queries from Google DeepMind for precise evaluation of LLM factual knowledge.

  • Question answering
  • English
9.4K 561K–10K rows
DatasetAI4Math/MathVerse

MathVerse: visual math problems

Visual math problem benchmark for multimodal LLMs.

  • Multiple choice
  • Question answering
  • Visual question answering
  • English
9.3K 731K–10K rows
Datasettheblackcat102/evol-codealpaca-v1

English GPT-4 coding instructions

Programming instruction-response pairs augmented by GPT-4 for training code assistants.

  • Text generation
  • English
9.3K 194100K–1M rows
Datasetb-mc2/sql-create-context

Text-to-SQL with CREATE TABLE context

78,577 examples of natural language queries, SQL CREATE TABLE, and SQL queries for text-to-SQL.

  • Text generation
  • Question answering
  • Table question answering
  • English
9.3K 51810K–100K rows
Datasetcommon-pile/stackexchange

Large Stack Exchange Q&A corpus

Tens of millions of English discussions from various Stack Exchange communities for text generation training.

  • Text generation
  • English
9.2K 1110M–100M rows
Datasetairtrain-ai/fineweb-edu-fortified

Cleaned FineWeb-Edu with embeddings

Large English FineWeb-Edu corpus after precise deduplication, featuring embeddings and row repetition counters.

  • Text generation
  • English
9.2K 71100M–1B rows
Datasetallenai/dolma3_mix-150B-1025Top pick

Dolma 3 mixed pre-training dataset

150B-token training data sample of Dolma 3 from web, code, math, and scientific documents.

  • Text generation
  • English
9.2K 1210M–100M rows
Datasetopendatalab/AICCTop pick

AI-ready multilingual Common Crawl

Large multilingual corpus from Common Crawl, cleaned for training language models.

  • Text generation
  • Multilingual
9.2K 1171B–10B rows
Datasetknkarthick/dialogsum

DialogSum: dialogue summarization

13,460 English dialogues with summaries for training and evaluating models.

  • Summarization
  • Text generation
  • English
9.2K 25910K–100K rows
Datasetopen-thoughts/AgentTroveTop pick

Agent interaction trajectories

1.7M open agent trajectories from 219 datasets: code, shell, math, and computer use.

  • Text generation
  • English
9.1K 2071M–10M rows
DatasetHelsinki-NLP/opus_books

OPUS parallel books for translation

Aligned books in multiple languages for training and evaluating translation models.

  • Translation
  • Catalan
  • German
  • Greek
  • +9
9K 1021M–10M rows
Dataseteaddario/imatrix-calibration

Imatrix calibration dataset for LLaMA.cpp

Small text dataset for calculating the importance matrix during model quantization.

  • Text generation
  • Arabic
  • German
  • English
  • +9
9K 79up to 1K rows
DatasetTIGER-Lab/WebInstructSub

Web instructions from forums

High-quality instruction subset from web forums for LLM training.

  • Question answering
  • English
8.9K 1731M–10M rows
Datasetbitext/Bitext-customer-support-llm-chatbot-training-dataset

Synthetic customer support dialogues

Annotated English support inquiries for fine-tuning LLM chatbots.

  • Question answering
  • Table question answering
  • English
8.9K 21010K–100K rows
Datasetnvidia/Nemotron-Personas-Vietnam

Synthetic personas of Vietnam

Synthetic personas based on demographics, geography, and personality traits of Vietnamese residents.

  • Text generation
  • Vietnamese
8.8K 2471M–10M rows
DatasetZyphra/ZydaTop pick

1.3T corpus for training language models

Large English open dataset corpus with filtering and deduplication for LM training.

  • Text generation
  • English
8.7K 851B–10B rows
DatasetSynthLabsAI/Big-Math-RL-VerifiedTop pick

250k verified math problems

Large dataset of math problems filtered for RL training of language models.

  • Question answering
  • Text generation
  • English
8.6K 245100K–1M rows🔒 on request
DatasetMegaScience/TextbookReasoning

Natural science texts for reasoning

A large English-language natural science dataset for post-training models with logical reasoning skills.

  • Text generation
  • English
8.5K 37100K–1M rows