1 paper · 1 filter
Adrienne Deganutti, Purvanshi Mehta, Simon Hadfield +1
Text-to-image models excel at natural image synthesis but struggle with graphic design, where success depends on satisfying precise constraints on typography, layout, color, and vi…