1 paper
Xuanyi Liu, Deyi Ji, Junyu Lu +6
Despite the impressive capabilities of text-to-image (T2I) models, an intent-generation gap often persists due to the brevity and ambiguity of user prompts. Existing approaches pri…