Loading...
Loading...
Image-text pairs, instruction tuning, visual QA, cross-modal data, foundation model training data
1,956 datasets
Zaynoid published a dataset titled 'Vlm Train 1K' on the Hugging Face platform on 2025-12-14. The title suggests it is likely a collection of 1,000 items for training vision-language models. The specific content, format, and structure require verification after download.
EndoVQA-Instruct is a multi-modal dataset containing endoscopy images and associated text, designed for benchmarking multi-modal large language models in medical analysis. The dataset includes images from the in-house WCE2025 collection and is managed by author Saint-lsy. Access to the data is restricted and requires formal request and approval.
From 1989 onwards, this dataset contains multimodal analyses of front pages from two Spanish newspapers, El País and ABC, covering the fall of the Berlin Wall. It was created by Silvia Molina Plaza and focuses on layout structure and rhetorical argumentation. The dataset was last updated on October 14, 2025.
EditReward-Data is a large-scale human preference dataset for instruction-guided image editing, introduced by TIGER-Lab. It comprises over 200,000 manually annotated preference pairs curated by trained experts following a standardized protocol. The dataset was last updated on October 12, 2025.
Published on huggingface by author zaiquan and last updated on 2025-12-04. The dataset likely contains multimodal data for spatio-temporal video grounding tasks, which involve linking language queries to specific objects and time segments in videos. Its specific content, scale, and collection methodology require verification after download.
A multimodal spiritual dataset featuring Ikaros in Spanish, Jiv Jago in Hindi, and languages of Russia, CIS, and Ukraine. The dataset was created by nativemind and was last updated on October 24, 2025. It includes enhanced multimodal data and supports up to 50 language examples.
ConstructionSite 10k contains 10,013 construction site images and annotations released by LouisChen15 in October 2025. The collection is partitioned into 7,009 training and 3,004 test samples specifically designed to evaluate Vision Language Models (VLMs) in civil engineering contexts.
Relation252K contains source-target image pairs across 218 distinct image editing tasks, released by handsomeWilliam in 2025. It serves as the evaluation set for the RelationAdapter model, focusing on the transfer of visual relations within Diffusion Transformers.
Datasetloom is an open-source platform for constructing and evaluating datasets for multimodal large language models (VLMs), developed by 599yongyang and updated in December 2025. It provides a full-stack framework using TypeScript, Next.js, and NestJS to streamline the creation of training data for vision-language tasks.
EditScore provides a series of open-source reward models ranging from 7B to 72B parameters for evaluating instruction-guided image editing. The benchmark likely contains data used to train and evaluate these models, with the largest model reportedly surpassing GPT-5 on their internal benchmark. The dataset was last updated on October 17, 2025.
A Visual Question Answering training dataset compiled from ChartQA, InfographicVQA, and MMLongBench-Doc. The dataset was created by openbmb and last updated on October 14, 2025. It appears to contain image data paired with text, likely for training multimodal models.
73 newsbites from eight major European newspapers published in the three days following the January 8, 2023, attack on Brazil's federal government buildings. Isabel Alonso Belmonte collected this multilingual sample to explore the multimodal construction of the political event. The dataset was last updated on October 14, 2025.
The dataset supports the 2026 Soccernet Challenge for multimodal (text, image, video) multiple-choice question answering. It covers 14 distinct soccer understanding tasks, including assessing player and team background knowledge, determining camera status, classifying actions, and recognizing fouls. The dataset was created by SoccerNet and last updated in October 2025.
Robo2VLM 1 provides between 100,000 and 1,000,000 visual question-answering records derived from real-world robot manipulation trajectories. Created by researcher keplerccc and updated in late 2025, the dataset uses multi-modal robot data to enhance scene understanding in vision-language models. It bridges the gap between internet-scale image-text corpora and specific robotic visuomotor policies.
RLAIF-V provides between 10,000 and 100,000 multimodal preference-alignment records developed by OpenBMB to improve Multimodal Large Language Model (MLLM) trustworthiness. The data utilizes AI-generated feedback to refine model responses, serving as a core training component for the MiniCPM-V 4.5 model released in 2024.
ClaraVid is a synthetic dataset for semantic and geometric neural reconstruction from low altitude UAV imagery. It contains 16,917 multimodal frames collected across 8 UAV missions over diverse environments. The dataset was created by radubeche and was last updated on October 31, 2025.
PKU-Alignment developed this dataset to facilitate Constrained Value Alignment through Safe Reinforcement Learning from Human Feedback (Safe RLHF). It provides human-annotated preference data for Large Language Models, specifically targeting the balance between helpfulness and safety constraints as of late 2024.
3,192 image–annotation pairs for evaluating vision-language models on Chinese calligraphy. The dataset supports tasks like full-page recognition and contextual visual question answering, including author identification and bilingual interpretation. It was created by author gtang666 and last updated on Hugging Face in October 2025.
40,000 samples with five strictly aligned modalities provide foundational data for AI systems to interpret CAD drawings. The dataset, created by jackluoluo and last updated in October 2025, is designed to address the challenge of understanding and utilizing computer-aided design data.
LoRA checkpoints were tuned on common sense reasoning datasets using a 0.5 billion parameter foundation model. The checkpoints are served as training data for DnD. The repository was created by Jerrylz and last updated on November 21, 2025.