Skip to content

Dataset · howard-hou

Visual question answering dataset with OCR

howard-hou/OCR-VQA

Hundreds of thousands of image-text examples for OCR and visual question answering tasks.

What's inside

Contains image and text data totaling between 100,000 and 1 million examples. The repository card is empty, so the language, exact field structure, and data source are not specified.

What you can do with it

  • Training multimodal models to recognize text in images (OCR)
  • Training visual question answering (VQA) systems on images containing text
Size
100K–1M rows
Downloads this month
2.2K
Likes
60

How to load it

from datasets import load_dataset

ds = load_dataset("howard-hou/OCR-VQA")