This wildcard list provides 500 unique women's undergarment names.
This is to assist image generation models such as Fooocus and its derivatives apply random keywords within an prompt. It should be saved as a .txt file in the wildcards directory. (Note that Ruined Fooocus uses a different location.)
Usage: Simply put two underscores around the filename prefix. i.e. __undergarments__ in the prompt. For example "1girl, 21 y.o. wearing __undergarments__"
This will then cause one of the items in the list to be inserted into the prompt. For example "1girl, 21 y.o. wearing satin tap shorts", next prompt in the batch could be "1girl, 21 y.o. wearing bodysuit".
Please be aware that 'nothing' is one of the keywords. Also keep in mind that the checkpoint or lora may not recognize the specific type of clothing and so leave the lady bereft of any coverage.
Description
Initial version. Contains the word 'nothing' so may produce nude results
FAQ
Comments (7)
noob here, Does this work in comfyUI? Do I need to install foooocus first? You do not have to explain the process of installing it, i'll find that on youtube just a simple yes or no is fine
Hi, Yes, this should also work in comfyUI. The resource, undergarments.txt, is just a basic text file containing a list of different clothes. You'll need to look at how to implement wildcards in comfyUI though. I have no experience there, but I'm sure Google, ChatGPT or YouTube will have some information, if not this site itself.
I found some information on how to get wildcards running on ComfyUI: https://ericri.medium.com/stable-diffusion-sdxl-wildcards-and-comfyui-e8483e00e1c
@swedishViking You went above and beyond, thank you, I will try and let you know
@swedishViking It is not as easy as the document says, you need some custom python nodes like show txt and change the clip text encode to getting that text input, I'll put an image in this thread, it should have the workflow if you ever decide to use comfyUI, no pressure just if you are curious
@serget2 Thanks. I'm using WebUI Forge 99% of the time, but if I find a compelling reason to switch to Comfy UI I will. From an 'outsiders' perspective it looks overly complex, but undoubtedly that also comes with some additional flexibility.
@swedishViking Well like I said I am a noob, I started with Automatic 1.1.1.1 and like a moth to a flame I started consuming all there is on Stable diffusion, that is how I got to ComfyUI, Yes one the one hand it looks complex, but the experience I have had with it so far (like 2 weeks), it is pretty positive, download a PNG and load it and you get the workflow to make it, press "que prompt" and it will tell you what you are missing, Models, loras, VAE's upscalers, specialized nodes.....and if you do not connect the wires it will tell you that aswell, and they are color coded it will not let you connect the wrong colors, So I would say I got the basics in like 2 days, now I am looking for more complex workflows from really good images, and a bunch of them are JPEG so bummer No workflow, but then I just build the workflow with what I know. Anyway like I also said there is no pressure from me forcing you to use ComfyUI, you seem to be very well skilled in webUI forge so if that works for you keep at it