/wit

WIT (Wikipedia-based Image Text) Dataset is a large multimodal multilingual dataset comprising 37M+ image-text sets with 11M+ unique images across 100+ languages.

OtherNOASSERTION

Watchers

No one’s watching this repository yet.