The code and models for the paper: Switti: Designing Scale-Wise Transformers for Text-to-Image Synthesis - View it on GitHub