Hi, I am doing some error analysis on my MMT system (follow your paper ).
But for images in test_flickr_2016, I can locate them as one of the 31,783 flickr30k images.
However, for test_filckr_2017 dataset, I couldn't. Those image names seem quite different from others (e.g., from train, val, test_flicker_2016).
Where could I find corresponding images for those in test_filckr_2017?
Hi, I am doing some error analysis on my MMT system (follow your paper).
But for images in test_flickr_2016, I can locate them as one of the 31,783 flickr30k images.
However, for test_filckr_2017 dataset, I couldn't. Those image names seem quite different from others (e.g., from train, val, test_flicker_2016).
Where could I find corresponding images for those in test_filckr_2017?