Hi, I am struggling with text prompts for multiple persons. For example if I want one in a dress and the other in a suit, it either mixes them together for both or just ignores one. Is there any trick to give specific instructions for each character in the image?
This is definitely a common problem. If you want you can get regional prompter, from https://github.com/hako-mikan/sd-webui-regional-prompter/, which exists exactly to circumvent this problem.
Edit to add the tutorial from stable-diffusion-art, which is a quite nice explanation: https://stable-diffusion-art.com/regional-prompter/
You can draw in areas and merge the whole at the end.