🎞 LoRA is trained purely on anime screencaps
🗝 Activation tags are in the right panel
📝 Note:
WARNING! This LoRA does not guarantee the text on the shirt will be 100% accurate.
Note: Pray to the AI RNG God for the right kanji characters (katakana seems a bit more stable lol). I had the best results at around 0.6-0.8 weight. Higher weight starts to mess up the style Needs you to experiment a bit with values to get it somewhat stable :)
As this is XL and not Flux, it doesn't understand letters. So this is the best I can do.
📸 Always use ADetailer to get detailed eyes
🛠 Technical notes:
LoRA shoud be compatible with Illu models & NoobAI. (Not tested.)
LoRA is compatible with various styles - Lower the weight & remove "anime screencap" from the prompt, if you are getting too strong anime style
Hi. res fix used on all sample images:
Upscaler: R-ESRGAN 4x+ Anime6B
Steps: 10-15
Denoising: 0.3
📬Leave a like, comment, review but first of all, ENJOY ;)