Skip to content

Dataset · apple

Apple synthetic captions and embeddings for DataComp-1B

apple/DataCompDR-1B

Apple's extension of the DataComp-1B dataset with synthetic captions, embeddings, and metadata for image-text pairs.

What's inside

Apple's dataset is an improved version of DataComp-1B ranging from 1 to 10 billion English records. It contains synthetic text descriptions (captions), embeddings, and metadata generated by pre-trained multimodal models on top of the original images and texts.

What you can do with it

  • Training CLIP-class multimodal models and image-text matching systems
  • Training text-to-image generation and image captioning models
Tasks
Text → image, Image → text
Languages
English
Size
1B–10B rows
Downloads this month
29.4K
Likes
37

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("apple/DataCompDR-1B")