2 datasets found

The dataset used in the paper is the MSLR-WEB30K dataset and the Yahoo! LTR dataset, which are the largest public LTR datasets from commercial search engines.

A dataset for evaluating the performance of image description models.