1 paper
Mahmoud Ahmed, Omer Moussa, Ismail Shaheen +5
One of the major challenges in training deep neural networks for text-to-image generation is the significant linguistic discrepancy between ground-truth captions of each image in m…