Kolors

Effective Training of Diffusion Model for Photorealistic Text-to-Image Synthesis