Geometric image-text VQA pairs
A VQA dataset with geometric diagram images and English text question-answer pairs, containing about 1 to 10k examples.
- Visual question answering
- English
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
A VQA dataset with geometric diagram images and English text question-answer pairs, containing about 1 to 10k examples.
Dataset with text, images, audio, and tables for multimodal models (100K-1M entries).
Official Chinese-language dataset for multi-condition text and visual product retrieval in e-commerce.
A small test sample of the MedHEB-Bench benchmark for evaluating multimodal models.
About 25.5 thousand pairs of scanned book pages and human-verified Telugu text from Wikisource.
A small sample of two-way conversations with separated speaker tracks for ASR.
100K to 1M question-answer pairs with reasoning over text-dense images across 5 European languages.
A cybersecurity corpus for training and question-answering tasks in English.
An evaluation set for LLM code generation consisting of HumanEval and MBPP translated into 22 programming languages.
33,889 electrical schematic images, text descriptions, and SPICE files collected from 62 textbooks.
735K synthetic reasoning examples across 28K competitive programming tasks.
Difficulty-filtered executable software development tasks for training and evaluating agentic models via RL.
NVIDIA post-training SFT and RL data for math, code, reasoning, and instructions across six languages.
Tabular and text registry of hourly Hugging Face model card checks for training data disclosures.
182k alternating text and image examples for step-by-step visual reasoning.
A massive archive of market data and trader behavior from a decentralized exchange.
English multiple-choice tests for evaluating models on dangerous domain knowledge and unlearning methods.
250 math problems from GSM8K translated by humans into 10 languages, including Russian, to evaluate reasoning.
Session logs of the Swival agent performing security vulnerability audits on open-source software.
Declassified U.S. National Archives PDF documents concerning the JFK assassination, digitized in Markdown.
A large dataset (100K-1M records) with video, sensor readings, and clothing folding annotations in real homes.
Apple's extension of the DataComp-1B dataset with synthetic captions, embeddings, and metadata for image-text pairs.
Unified multimodal dataset combining images, audio, and text from public sources.
64,878 English QA pairs for 40,263 spine images to evaluate multimodal models.