This is my collection of useful Wildcards. I created most of them using ChatGPT to streamline my work with Stable Diffusion and avoid using the Dynamic Prompts extension, which I used before. Through my experience, I've found that Wildcards are more convenient to use.
The purpose of these Wildcards is to introduce randomness and diversity into the created graphics. They are mainly intended to help you in creating scenes with people.
To use these Wildcards, you need an extension. You can use the official one stable-diffusion-webui-wildcards.
Of course, there are other extensions that allow you to use Wildcards. I recommend the one I mentioned because it's the simplest.
Most likely, after installing the extension, you'll need to restart Stable Diffusion (or the rented server instance) for the extension to work correctly. A simple reload of WebUI may not be sufficient.
Once you have a working Wildcards extension, you just need to copy wildcard files to the appropriate directory.
In the case of stable-diffusion-webui-wildcards, it will be stable-diffusion-webui/extensions/stable-diffusion-webui-wildcards/wildcards/
- You can manually download these files as a compressed archive and extract them by yourself.
- You can download them using git-clone (output directory must be empty).
Let's say you want to generate a scene with a woman in a random location. Let her clothing be random as well.
photography of __nationalities__ woman BREAK wearing __colors__ __clothes_upper__ BREAK in __locations__
The initial prompt will look as follows:
photography of Polish woman BREAK wearing purple jacket BREAK in city street
If you believe something is missing, that something could be useful, or that something should be removed, go ahead - fork this repository, edit the files, and submit a pull request.
- The list of artists was obtained from the Stable Diffusion Cheat-Sheet website.