Style Transfer?

[oivio]

There is so much in this amazing webuiand maybe I just could not understand or find it but, is there a way to have a clean option for style transfer based on another image?
As an example:

Some Ref: https://old.reddit.com/r/StableDiffusion/comments/wvzr7s/tutorial_fine_tuning_stable_diffusion_using_only/

[ClashSAN]

this is already supported, this is called textual inversion. It uses .bin or .pt file which you can grab from https://cyberes.github.io/stable-diffusion-textual-inversion-models/
make a folder called "embeddings" root directory, place your .bin or .pt files in, and use the name of the file in your prompt: "flowers in the style of nebula", with nebula.bin or nebula.pt as the file

[Filarius]

Problem is about you need at last 10 GB VRAM to make your own, and 20+ GB to make it resonably fast.
After you made training this Web-UI can use saved file to add into SD.

[TheDaoOfCode]

I downloaded hugging face concept/embedding but am I'm getting an error in the command promt: "Error loading emedding moebius.bin" ..this happens for all the bin files, however when using the keyword the style seem to work. Just for clarification - is the error and error? or is the style coming from vanilla SD ? Any advice much appreciated.

[przemoc]

Today AUTOMATIC1111 created page on wiki:
https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Textual-Inversion

[m3ngyang]

Does the Textual Inversion work in the img2img? Thx.

[missionfloyd]

This thread is seven months old. Nowadays, controlnet is used for style transfer.

https://www.youtube.com/watch?v=PbDdtPTYm_4