Region-Aware Text-to-Image Generation via Hard Binding and Soft Refinement

 

Note: A simplified demo of our RAG-Diffusion (For more complex layouts, please run our code directly).