Fhrozen/flickr30k-narratives

Flick30K Narratives数据集是一个图像和文本配对的数据集，用于图像到文本的生成任务。它包含了来自原始Flickr30K的图像和注释，以及localized-narratives项目的注释。数据集的语言为英文，大小在10K到100K之间。Flick30K Narratives dataset is a collection of image-text pairs for image-to-text generation tasks. It includes images and annotations from the original Flickr30K, as well as annotations from the localized-narratives project. The dataset is in English and its size ranges from 10K to 100K.