🎞 LoRA is trained purely on anime screencaps
🗝 Activation tags are in the right panel
📝 Note: I wanted a bit more episodes to release, to get the armor outfit more stable with more images, but it should work somewhat okay :) That "pelvic curtain" is pain in the ass tho.
📸 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 ;)
Description
FAQ
Comments (12)
Himenohara Suzuran next
It doesn't work for me in comfyui
I don't work with comfy, I'm using reforge and it works perfectly fine. You'll have to ask someone else for help as there's no reason it would not work in both.
People don't appreciate your loras enough! This is the only cha-hae lora that worked well for me and it's even flexible with a lot of styles I throw at it. Keep up the great work ❤️
Thank you for this amazing LoRA model!
What kind of model did you use for ADetailer? And did you add any LoRA for the eyes?
yeah im wondering this too about the ADetailer
@leonisterb @m0n3rch I'm using basic model for detection: face_yolov8n.pt
However, lately I've been told that Anhzc20 face and eye detection are much better as they are not square but detect directly just face without hair.
No additional LoRA has been used for the eyes. Empty prompt in the adetailer.
@LittleJelly Ohh, I see, thanks!
This is one best character of this anime
your models for solo leveling are gorgeous.
Your work so good!
may you do the Japanese girls and the arise girl as well?
I've retired from LoRA making, sorry nothing new coming from me anytime soon.