This is a great textual inversion made on the basis of high-quality images.
Works great on most popular Stable Diffusion 1.5 models.
Tested image generation on different styles and scales.
Watch your negative promt, some embeddings can suppress the result. Then you will either not get a believable image, or it will be monotonous - regardless of positive tokens.
I work on many interesting things!
❤ You can support me on ko-fi.com ☕
Description
FAQ
Comments (2)
The example images you've included look nothing like her to the point where, if you hadn't labled the TI as Ana de Armas, I'd have no idea they were meant to be her.
I suggest strongly to Never use a checkpoint which is not meant for creating Photo-Realistic images.
Otherwise, the faces will have distortions and will not be the same as the face the model was trained on.
There is a HUGE difference between Anime-style checkpoints and Photography-realistic picture Checkpoints.
- For Photo-Realistic images, there is a proper way to write a prompt if you want to get the Best results from your A.I.
- NEVER use the person Real Name on a trained model. There are many reasons why this should never be done.
If you are interested in knowing some S.D. AI secrets, I can help you.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.