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

Datasetnvidia/Nemotron-Pretraining-Indic-CC-v1Top pick

Synthetic corpus in 22 Indian languages from NVIDIA

Huge text corpus of 1 to 10 billion records in 22 Indian languages for language model pre-training.

  • Text generation
  • Assamese
  • Bangla
  • Bodo
  • +9
5.5K 61B–10B rows🔒 on request
Datasetnvidia/Nemotron-Pretraining-Code-v2Top pick

Nemotron code and math pre-training corpus

NVIDIA's updated pre-training corpus with code, math, English web text, and synthetics.

  • Text generation
5.5K 139100M–1B rows🔒 on request
Datasetarmand0e/claude-fable-5-claude-code

Anonymized Claude Fable 5 traces

Raw anonymized traces of the Claude Fable 5 agent for training and analyzing agent scenarios.

  • Text generation
5.4K 430up to 1K rows
Datasettucnguyen/ShareChat

Real conversations with popular AI chatbots

Over 142 thousand live chats with ChatGPT, Claude, Gemini, Grok, and Perplexity for chatbot training and prompt analysis.

  • Text generation
  • Question answering
  • English
  • Multilingual
5.4K 531M–10M rows🔒 on request
DatasetTHUIR/Qilin

Multimodal user session dataset for RAG and search

Large Chinese dataset (1M to 10M) for search, recommendations, and RAG systems.

  • Question answering
  • Text classification
  • Sentence similarity
  • Chinese
5.3K 441M–10M rows
DatasetUCSC-VLAA/MedReason

Medical reasoning for LLMs

A large set of medical tasks with step-by-step reasoning for explainable LLM answers.

  • Question answering
5.2K 9010K–100K rows
Datasettokyotech-llm/swallow-code

SwallowCode: EN/JA code corpus

Large English and Japanese code corpus for training and fine-tuning language models.

  • Text generation
  • English
  • Japanese
5.2K 84100M–1B rows
DatasetOpenMed/Medical-Reasoning-SFT-GPT-OSS-120B

Medical dialogues with reasoning for SFT

Synthetic medical dialogues with reasoning chains for SFT of LLMs in healthcare.

  • Text generation
  • English
5.2K 267100K–1M rows
DatasetHuggingFaceH4/CodeAlpaca_20K

CodeAlpaca 20K: train and test split

Original CodeAlpaca split into train and test; about 20k text examples for code generation.

  • Text generation
5.1K 11610K–100K rows
DatasetAnthropic/model-written-evals

Evaluations written by language models

A collection of eval tasks generated by LLMs to test model behavior.

  • Multiple choice
  • Zero shot classification
  • Question answering
  • English
5K 711K–10K rows
Dataset0xDing/wikipedia-cn-20230720-filtered

Filtered Chinese Wikipedia 2023

254k cleaned Chinese Wikipedia articles for training generative models in Chinese.

  • Text generation
  • Chinese
4.9K 182100K–1M rows
DatasetWe-Math/We-Math

We-Math visual math problem benchmark (ACL 2025)

6.5k visual math problems with hierarchical knowledge annotation for evaluating VLM reasoning.

  • Question answering
  • Text generation
  • English
4.9K 351K–10K rows
Datasetnvidia/GranaryTop pick

Granary: speech and translation in 25 languages

A large open speech dataset across 25 European languages for ASR and speech translation.

  • Automatic speech recognition
  • Translation
  • Bulgarian
  • Czech
  • Danish
  • +9
4.8K 238100M–1B rows
Datasetbakrianoo/jabarti-llm-dataset

Corpus for small LLM: Egypt and Wikipedia

Clean sectioned corpus in Arabic and English for pre-training a small LLM.

  • Text generation
  • Question answering
  • Arabic
  • English
4.7K 1651M–10M rows
Datasetnvidia/Nemotron-Pretraining-Specialized-v1.1Top pick

Specialized synthetic data from NVIDIA

Large corpus of synthetic texts on code, logic, and economics for LLM pre-training.

  • Text generation
4.7K 4510M–100M rows
DatasetSalesforce/APIGen-MT-5k

5000 multi-step agent trajectories

5000 multi-step agent trajectories for training and evaluating agent models in English.

  • Question answering
  • Text generation
  • Reinforcement learning
  • English
4.7K 1181K–10K rows
DatasetCongliu/Chinese-DeepSeek-R1-Distill-data-110k-SFT

Chinese DeepSeek R1 distillation SFT dataset

110k Chinese examples with R1 answers gathered in the output field for SFT training.

  • Text generation
  • Question answering
  • Chinese
4.7K 233100K–1M rows
Datasetccdv/pubmed-summarization

PubMed article summarization

PubMed corpus for summarizing long medical articles: article and abstract.

  • Summarization
  • Text generation
  • English
4.7K 100100K–1M rows
Datasetai4privacy/pii-masking-200k

PII masking on 200k examples

Texts with personally identifiable information in English, French, German, and Italian for privacy masking training.

  • Text classification
  • Token classification
  • Table question answering
  • English
  • French
  • German
  • +1
4.7K 132100K–1M rows