MMID Image Classification Dataset

The MMID dataset is a large-scale multilingual image dataset released by the University of Pennsylvania. The dataset is doubly parallel: for each language, words are stored in parallel with images representing the words, and in parallel with the translation of the words into English (and the corresponding images). Currently, the dataset has 98 languages, with a maximum of 10,000 words per language. This download only provides the Chinese version. For details, please visit:http://multilingual-images.org/