The latest resource contains 112 embeddings for different female faces that you can use royalty-free for any SD 1.5 model. (My goal is to roll out 100 200+ embeddings for different female faces, a SDXL LoRA, and a Flux LoRA with the same faces as a solution to the "same-face phenomenon".)
Please place them in the appropriate embeddings folder, e.g. stable-diffusion-webui/embeddings.
If you're using Dynamic Wildcards, please place the virtualgoddess.txt file in the wildcard folder: stable-diffusion-webui/extensions/sd-dynamic-prompts/wildcards.
After you've completed this setup, you can go into Automatic1111 and put __virtualgoddess__ in your prompt and pick one of these faces at random.
If you want even more faces, you can put the following in your prompt: [__virtualgoddess__:__virtualgoddess__] which mixes one random face with another.
Note, if you're having trouble opening any of the .pt files in Automatic1111, please use version 89 and higher.
Recommended Base Models:
Description
FAQ
Comments (4)
hai, the zipped file somehow is password-locked when tried to extract
My bad, fixed now
Thanks for that tip on mixing. I hadn't yet seen that anywhere before. Any chance you've found a resource that covers all useful syntaxes like that? I've only found parts/pieces.
Yeah, there's a bunch on the Automatic1111 features page, this one is called "Prompt editing" (scroll down): https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Features
I've only touched the tip of the iceberg on a lot of these new features...