Skip to content

Dataset · OpenFace-CQUPT

15 million face and text caption pairs

OpenFace-CQUPT/FaceCaption-15M

A large English dataset of facial images with captions for description and generation.

What's inside

FaceCaption-15M is a large English dataset of facial image and text description pairs. To obtain the images, users must download the data and separately request access to the source Laion-face per agreement. The card features diverse and high-quality facial images with natural captions.

What you can do with it

  • Train image-to-text facial description models
  • Fine-tune text-to-image LoRA on portraits
  • Evaluate facial generation quality
Tasks
Image → text, Text → image
Languages
English
Size
10M–100M rows
Downloads this month
223
Likes
107

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("OpenFace-CQUPT/FaceCaption-15M")