50K hours of first-person hand action videos
Large multimodal corpus of human hand manipulation videos with motion annotations and descriptions for robot learning.
- Robotics
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.
Answered by Gemini 4 · the chat is kept only in this tab
Large multimodal corpus of human hand manipulation videos with motion annotations and descriptions for robot learning.
Collection of tasks and data for training and evaluating English text understanding systems.
1.3T tokens of filtered educational web pages in English for LLM training.
Large-scale multilingual archive of news articles from 2016–2026 with quality filtering and deduplication.
Large English web corpus broken down by dozens of domains and iteration statistics for training and classification.
12k complex multi-discipline questions for rigorous LLM evaluation.
Top 10% highest-quality and structured web texts from the FineWeb corpus for language model training.
Over 18.5 trillion tokens of cleaned and deduplicated English web texts from CommonCrawl for LLM training.
Millions of records with language model rating history from the popular Arena benchmark.
English and German text scripts with sudden emotional shifts for a single character to train expressive TTS models.
Large-scale English corpus of 1.61 million web novel records containing text, tables, and images.
Over 10,000 hours of real-world recordings of craftsmanship, cooking, and construction work with IMU sensor data.
Over 1.1 million hours of web audio in 100+ languages for ASR, TTS, and audio representation learning.
4 trillion tokens and 3 billion documents for LLM pre-training; strong benchmark results.
A massive dataset of over 1.1 million hours of audio across 100+ languages for speech tasks.
Over 10 thousand Arabic books in PDF, TXT, and DOCX formats, recognized via Google Document AI.
FineWeb-2 pretraining corpus with 1000+ languages for training language models.
Challenging NLU tasks including textual entailment, QA, and coreference resolution for language model evaluation.
An open multilingual corpus of 2.27 trillion tokens spanning books, press, science, law, and code.
Unedited daily and working first-person video recordings in English for training video models.
English synthetic text data generated by SmolLM2 from the FineWeb-Edu corpus.
Corpus of over 4.6 thousand Arabic books (over 5 million pages) with extracted text.
Multimodal dataset with video, 3D hand and object poses in real-world conditions for CVPR 2026.
English QA benchmark with questions requiring multi-step reasoning and general knowledge.