Skip to content

Dataset · yushaohan

Interleaved Text and Image Generation Dataset

yushaohan/IntReal-100K

10k to 100k real multimodal sequences of interleaved text and images with instructions.

What's inside

This English-language corpus contains 10k to 100k examples for interleaved text and image generation. Covering four categories—storytelling, knowledge, planning, and tutorials—the data includes task instructions, with the test split augmented by detailed evaluation rubrics.

What you can do with it

  • Supervised fine-tuning of multimodal models capable of jointly generating text and illustrations
  • Building systems for automated layout of illustrated instructions and stories
  • Evaluating the quality of multimodal generation using built-in rubrics
Tasks
Multimodal
Languages
English
Size
10K–100K rows
Downloads this month
3.3K
Likes
4

How to load it

from datasets import load_dataset

ds = load_dataset("yushaohan/IntReal-100K")