---
annotations_creators: []
language: []
language_creators: []
license: []
multilinguality: []
pretty_name: 'Internet Archive historic book pages unlabelled.'
size_categories: []
source_datasets: []
tags: []
task_categories: []
task_ids: []
---

# Data card for Internet Archive historic book pages unlabelled.


- `10,844,387` unlabelled pages from historical books from the internet archive.
- Intended to be used for:
  - pre-training computer vision models in an unsupervised manner
  - using weak supervision to generate labels