Chart and diagram reasoning SFT dataset
1.8M chart QA pairs with Chain-of-Thought annotations for VLM training.
- Visual question answering
- Image text → text
- 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
1.8M chart QA pairs with Chain-of-Thought annotations for VLM training.
223 simulation-ready household, kitchen, and workshop objects in OpenUSD format with physics and semantics.
6.5k visual math problems with hierarchical knowledge annotation for evaluating VLM reasoning.
Photos of 30 diverse objects and pets from multiple angles for personalizing text-to-image models.
TAO dataset extension featuring amodal bounding boxes and segmentation masks for partially or fully hidden objects.
Image-text dataset from ICCV '25 for evaluating vision-language models in agriculture.
English geospatial dataset for multi-modal analysis and Earth imagery segmentation.
A classic English benchmark of 50,000 questions on document images for evaluating Document AI.
Tabular, text, and image data on 148,909 published Steam games for classification.
English multimodal dataset for training models to describe specific regions in images and videos in detail.
A dataset of 27,000 labeled Sentinel-2 RGB satellite images for land cover classification.
Fashion object detection dataset: images with segmentation masks and expert clothing ontology.
1 to 10 million generated Japanese user profiles considering real demographics for financial AI.
Page images from PubMed Central with layout block annotations for document structure recognition.
A high-resolution synthetic dataset with detailed text instructions for training image editing models.
Synthetic dataset of 100k-1M image-instruction pairs in English for training image editing models.
A dataset of ultra-high-resolution (UHR) images with captions for training high-detail image generation.
A classical dataset of individual handwritten text line images in English for OCR training.
25M+ medical images across 10 modalities with multimodal annotations for 65 diseases.
A large collection of 27 subsets featuring standard videos, egocentric recordings, and robot manipulation demonstrations.
Converted 3D models of 79 robots in .glb format with preserved joints for browsers and Blender.
Filtered and repaired 3D models from the defunct Google Poly library in GLB format.
Raw corpus of 2.3 million multimodal tasks and reasoning chains (8.8 billion tokens) before quality filtering.
Fewer than a thousand 3D models of diverse professionals and animals with a unified skeleton and facial animation.