IA_unlabelled / README.md
davanstrien's picture
davanstrien HF staff
Update README.md
dd7d748
---
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