Hugging Face
Models
Datasets
Spaces
Posts
Docs
Enterprise
Pricing
Log In
Sign Up
Datasets:
wikimedia
/
wit_base
like
59
Follow
Wikimedia
108
Tasks:
Image-to-Text
Text Retrieval
Modalities:
Image
Text
Formats:
parquet
Sub-tasks:
image-captioning
Languages:
Afrikaans
Aragonese
Arabic
+ 102
Size:
100K - 1M
ArXiv:
arxiv:
2103.01913
arxiv:
1512.03385
arxiv:
1905.00641
Tags:
text-image-retrieval
Libraries:
Datasets
Dask
Croissant
+ 1
License:
cc-by-sa-4.0
Dataset card
Viewer
Files
Files and versions
Community
5
7808b33
wit_base
/
data
/
train-00024-of-00330.parquet
Commit History
Upload data/train-00024-of-00330.parquet with git-lfs
2b79d70
Mario Šaško
commited on
May 2, 2022