Skip to content

Datasets

Hugging Face datasets with plain-language descriptions: what's inside and what models to train on them. 4,797 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

DatasetJr23xd23/ArabicText-LargeTop pick

Arabic corpus of 743k articles

743,288 Arabic articles, over 244 million words, a corpus for LLM training and fine-tuning.

  • Text generation
  • Fill mask
  • Text classification
  • Arabic
564 72100K–1M rows
Datasetsaidutta69/fable-5.1-premium-v2

Fable 5.1 agent dataset for SFT

SFT dataset based on 9,680 agent traces with tool schema support.

  • Text generation
  • English
561 61K–10K rows
DatasetAxiomicLabs/Tiny_Theory_of_Mind

Theory of mind benchmark for small language models

English test set for evaluating the ability of small base models to understand others' intentions and beliefs.

  • Question answering
  • English
550 491K–10K rows
Datasetsunzeyeah/chinese_chatgpt_corpus

Chinese corpus for SFT and RLHF

Chinese texts for supervised fine-tuning and RLHF in a ~5 GB JSONL file.

  • Text generation
  • Question answering
  • Reinforcement learning
  • Chinese
548 901M–10M rows
Datasetiamtarun/code_instructions_120k_alpaca

Code instructions in Alpaca format

120k programming instructions in Alpaca style with a prompt column for fine-tuning code models.

  • Text generation
  • Question answering
546 70100K–1M rows
Datasetmiromind-ai/MiroVerse-v0.1

MiroVerse v0.1 deep research data

English SFT and DPO data for training and reproducing the MiroThinker-v0.1 deep research agent on Qwen3.

  • Question answering
  • English
543 242100K–1M rows🔒 on request
DatasetLDJnr/Puffin

Multi-turn GPT-4 human dialogues

3,000 examples of GPT-4 responses across over 2,000 multi-turn English dialogues for chat model training.

  • Question answering
  • Text generation
  • English
542 971K–10K rows
DatasetOpenDataArena/ODA-Math-460k

ODA-Math-460k: math for training

460k cleaned English math reasoning examples for training and evaluating reasoning models.

  • Question answering
  • Text generation
  • English
539 106100K–1M rows
DatasetOpenLLM-France/Claire-Dialogue-French-0.1

French dialogues and theater plays

Corpus of live speech transcripts, debates, and plays in French for training conversational models.

  • Text generation
  • French
539 6110K–100K rows🔒 on request
DatasetOpenMOSS-Team/moss-002-sft-data

MOSS-002 dialog SFT data

An open dialog dataset for SFT: text-davinci-003 prompts and responses in English and Chinese.

  • Text generation
  • English
  • Chinese
533 961M–10M rows
Datasettrl-lab/SQaLe-2-text-to-SQL-QueriesTop pick

Large question-SQL query pair corpus for model training

Over 1.4 million English questions and 176k SQL queries based on real database schemas.

  • Text generation
  • Table question answering
  • English
530 6100K–1M rows
DatasetExploration-Lab/IL-TUR

Indian law benchmark across six languages

From 100K to 1M examples across 8 legal tasks in English and Indian languages.

  • Text classification
  • Summarization
  • Translation
  • English
  • Hindi
  • Bangla
  • +7
529 37100K–1M rows🔒 on request
Datasettrendmicro-ailab/Primus-Instruct

Cybersecurity instructions for LLMs

A small set of expert cybersecurity scenarios with GPT-4o responses for training specialized LLMs.

  • Text generation
  • English
529 22up to 1K rows🔒 on request
DatasetAiresPucrs/stanford-encyclopedia-philosophy

Stanford encyclopedia of philosophy texts

An English text corpus with a volume from 100 thousand to 1 million entries for generation and classification tasks.

  • Text classification
  • Text generation
  • English
528 56100K–1M rows
DatasetMathLLMs/MathCodeInstruct-Plus

Dataset for solving math problems via code

English text data for training LLMs in mathematical reasoning with programming code integration.

  • Question answering
  • Text generation
  • English
528 1910K–100K rows
Datasetianncity/GLM-5.2-ConversationTop pick

50,000 deep reasoning dialogues from GLM-5.2

A large 120M-token corpus of GLM-5.2 responses generated in high reasoning mode.

  • Text generation
  • Question answering
  • English
525 5910K–100K rows
DatasetKBlueLeaf/danbooru2023-metadata-database

Danbooru2023 Metadata in SQLite

An SQLite database of Danbooru2023 tags and posts with a latest ID of 7,866,491 for generative models.

  • Image classification
  • Text → image
  • Image → text
  • English
  • Japanese
522 841M–10M rows🔒 on request
DatasetNan-Do/instructional_code-search-net-python

Python code generation and description instructions

English instructional dataset for generating Python code from descriptions and explaining ready code.

  • Text generation
  • English
520 38100K–1M rows
DatasetMustafaege/qwen3.5-toolcalling-v2

SFT dataset for tool calling and Jupyter agents

Over 60 thousand English examples for training models on external tools and code in Qwen3 format.

  • Text generation
  • English
510 50100K–1M rows