Thank you all, you gave me enough buzz to start a new training session.
Training images were generated using the Uranami SD1.5 LoRA https://civitai.com/models/179849/uranami-azur-lane .
Version 1 was trained with 25 images I had on hand.
For version 1.1, I generated new images—still just 25, but more carefully selected.
Can force the traditional Japanese style for backgrounds. Check the release notes for more details.
The presentation images were made using Nanomix and Knk SoftShading as the refiner (switch 0.5).
Refer to the prompts of the example images if you encounter any usage issues.
LMAO the moderators are such terrified prudes. Look here for sample pictures: chalilinAI - pixiv
Description
created, beware I've constated about 1/3 generation contening errors:
Instability in the color of the kimono (blue decorations can turn red).
Instability in the fur (the tail and ears can turn reddish).
The face tends to lack accuracy. (compared to the cutness of the original art)
The ears often stick out of the hood. (like everytime, sadly)
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.