This is a 4-vector embedding of Billie Eilish made with a dataset of 90 images and 10000 steps.
I tested the embedding with different models. The sample images were generated with the following models.
It seems to perform particularly well with DreamShaper , Realistic Vision 1.2 and the different Protogen models.
It performs worse with Dreamlike Diffusion 1.0 and Analog Diffusion
at least according to the tests I've done.
Description
FAQ
Comments (5)
Sorry if this is a dumb question but where would I put this file to be able to use it?
If you use the automatic 1111, in your stable-diffusion-webui folder you should have an "embeddings" folder. You put the .pt file in and add the name of the embedding in your prompt. In this case it would be "BIL13L1SH". for example: "A picture of BIL13L1SH". Or you can use the example prompts I provided by pressing the exclamation point icon on the example images above. :)
@AngelicDiffusion Thank you so much! If I wanted to do something similar to this with another celeb is there an easy guide to do so?
@anonfaker it's going to be a bit complicated for me to explain this here, but personally I like the "Detailed guide on training embeddings on a person's likeness" on reddit or if you prefer video there is the "ULTIMATE FREE TEXTUAL INVERSION In Stable Diffusion! Your FACE INSIDE ALL MODELS!" on the Aitrepreneur channel. Finally you also have resources on the Civitai discords, I think there is a section for training models and textual inversion embedding.
This is great. one of the best i've seen. Curious if you do contract or I could fund some models.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.