metadata
license: creativeml-openrail-m
tags:
- text-to-image
- v2.0
- Embedding
Textual Inversion Embedding by ConflictX For SD 2.0 trained on 768x768 images from midjourney.
Install by downloading the step embedding you want, and put it in the \embeddings folder
It is slightly overfit on 150 steps so some concepts/keywords will be harder to prompt for (use negatives or weight Kipaki down) but it works amazing for cityscapes, people, gods, and other scifi genres. Very stylized on ancient Egypt, scifi, and orange/blue color scheme but other concepts are definitely possible: More images here: https://imgur.com/a/W2bmBaV
Use keyword: Kipaki-xxx
xxx is embedding number
There are multiple versions, the images below were created with the 150 step version.
Highres Images:
