🎞 LoRA is trained purely on anime screencaps
🗝 Activation tags are in the right panel
📸 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
fixed oversaturation / bloom
fixed washed out colors
removed bad images from the dataset
upscaled dataset images & got rid of blurriness
FAQ
Comments (3)
I notice that both versions have the same filename. Would you consider to add version numbers to it in the future for easier differentiation and management locally?
I would do it normally, however V1.0 had some really shitty images in the dataset. It's better to just forget about it and overwrite it with a new one.
amazing!