MMID 图像分类数据集

MMID 数据集是由宾夕法尼亚大学发布，是一个大规模多语言的图像数据集。并且数据集是双重平行的：对于每种语言，单词与表示单词的图像平行存储，并与单词翻译成英语（和相应的图像）平行。目前该数据集有 98 种语言，每种语言最多 10,000 个单词，此次下载仅提供中文版，详情可浏览：http://multilingual-images.org/

The MMID dataset, released by the University of Pennsylvania, is a large-scale multilingual image dataset. It features dual parallelism: for each language, each word is stored in parallel with the image that represents it, and also in parallel with its English translation along with the corresponding image for that translation. Currently, the dataset covers 98 languages, with up to 10,000 words per language. This download only provides the Chinese version, and for detailed information, please visit http://multilingual-images.org/

搜集汇总

背景与挑战

背景概述

MMID图像分类数据集由宾夕法尼亚大学发布，是一个大规模多语言图像数据集，具有双重平行结构，包含98种语言且每种语言最多10,000个单词。当前下载仅提供中文版本，数据大小为9.44 GB。