Skip to content

Dataset · common-canvas

Tens of millions of 4K photos with synthetic captions

common-canvas/commoncatalog-cc-by-nc

A large collection of high-resolution Yahoo Flickr images with English descriptions for generative models.

What's inside

An array ranging from 10M to 100M English image-text pairs. Images with resolutions up to 4K were collected in 2014 on Yahoo Flickr under various Creative Commons licenses. Synthetically generated text captions are added to the images.

What you can do with it

  • Training and fine-tuning high-resolution text-to-image generation models
  • Training multimodal encoders and visual content understanding systems
Tasks
Text → image
Languages
English
Size
10M–100M rows
Downloads this month
31.5K
Likes
8

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("common-canvas/commoncatalog-cc-by-nc")