Dataset Card for "AToMiC-All-Images_wi-pixels"
Languages
The dataset contains 108 languages in Wikipedia.
Data Instances
Each instance is an image, its representation in bytes, and its associated captions.
Intended Usage
- Image collection for Text-to-Image retrieval
- Image--Caption Retrieval/Generation/Translation
Licensing Information
CC BY-SA 4.0 international license
Citation Information
TBA
Acknowledgement
Thanks to: img2dataset Datasets
- Downloads last month
- 668
Homepage:
AToMiC homepage
Source:
WIT
Paper:
WIT: Wikipedia-based Image Text Dataset for Multimodal Multilingual Machine Learning