This model was trained using 3 photos of sitting cats in front view. [name] and [filetype] was used in the Prompt. The photos were renamed with respect to the usage of [filetype]. Base was a white cat with mechanical parts. Steps for training overall 10000.

This embedding has some problems when creating images, which I will briefly list here. The effect occurs that more than one cat appears even in a small picture. Denoising often leads to surreal images. In contrast to a few other embeddings of mine, you need a negative prompt for good pictures.

As you can see in the gallery, the Embedding model works with various Checkpoint models. The quality is always equally good.

To be compatible with the prompts in the gallery, the model must be called and saved as WhiteMechanicalCat.pt. With weighting, WhiteMechanicalCat:0.5 can then be used in the Prompt.

Nevertheless. Have fun. Be inspired!