How can I obtain the imDB and.npy or info.npy format provided by the link from the original COCO or Textcaps dataset? Because I want to optimize the feature extraction method of the original data set, but I know where I can get the processed imDB format data set provided by the paper. I am looking forward to your reply.
How can I obtain the imDB and.npy or info.npy format provided by the link from the original COCO or Textcaps dataset? Because I want to optimize the feature extraction method of the original data set, but I know where I can get the processed imDB format data set provided by the paper. I am looking forward to your reply.