1 paper · 1 filter
Gurusha Juneja, Sukrit Kumar
Diffusion models when conditioned on text prompts, generate realistic-looking images with intricate details. But most of these pre-trained models fail to generate accurate images w…