XTD10

ImagesTextsIntroduced 2021-09-15

XTD10 is a dataset for cross-lingual image retrieval and tagging consisting of the MSCOCO2014 caption test dataset annotated in 7 languages that were collected using a crowdsourcing platform.