Ratings are appreciated so i can see what i can improve on for my next models!
If you appreciate these models you can support me with a cup of coffee :) ->

Trained with AbyssOrangeMix2_hard.

The model consists of about 35 images and was trained in 1100 training steps.

Because there weren't many images available you might have to do a bit more finetuning to get the perfect results.

Recommended settings to get images similar to the examples:

Positive
(ultra-detailed), (masterpiece), best quality, high quality, 1girl, croissant arknights

LoRA strength: 0.6-0.8

Checkpoint + VAE: AOM2_Hard

Steps: 25

CFG: 7

Sampler: DPM++ 2M Karras

Feel free to reach out with any questions or suggestions you may have!