✨ Hello everyone! Welcome to another one of my LoRAs.
To use the original anime style in future in this and futures Loras, simply add these tags:
anime screencap, anime coloring, flat color,If the style isn't being fully captured, you can also include the anime's release decade tag, for example: 2020s (style)
In some cases If the model generates repetitive poses or expressions, use the tag:enh_cnr
You can adjust its strength with the following format: (enh_cnr:1.5). I do not recommend using a weight higher than 2.0.
💬 Suggest a Character
Have a favorite character? You can suggest it here. I review all proposals!
(But I can't promise to make them all :))
🎯 Commission ?
Of course you can requests any type of character or custom styles, contact me through:
Civitai: Send me a private message.
Ko-Fi: Cinabrio
Discord (Soon)
Each model is designed to:
Preserve the character's essence (face, body, unique features).
Offer freedom to mix outfits and styles.
Recommended Settings:
Sampler: Euler or DPM++ 2M (25–35 steps)
LoRA Weight: 0.8 (up to 1.0 for higher accuracy)
ADetailer: Enabled
CFG Scale: 5–6
Thank you for your support! Every download, comment, or Buzz⚡ helps me keep creating.
Enjoy generating with this LoRA!
Description
🧬Physical characteristics
shain_cnr, 1girl, short hair, purple hair, yellow eyes, small breasts,
👗Signature Outfit
armband,brown gloves, long sleeves, neck ribbon, underboob, white shirt,miniskirt, white strapless bra,brown skirt, unbuttoned shirt, untucked shirt,black thighhighs,white boots, jacket,
FAQ
Comments (3)
is that enh_cnr a default command/prompt or something YOU bake into your loras? I've never seen/heard of it before.
enh_cnr is a custom tag I incorporate into my LoRAs. It's used to manage the controlled noise that I integrate into the training process. This method allows me to make the model highly flexible, particularly with poses, expressions, and stylistic variations without altering or distorting the core character identity.
And yes, this is a feature specific to my LoRAs. Others may employ a similar principle, but this is my unique way of implementing and representing it. :)
@Cinabrio Fair enough.
Details
Available On (2 platforms)
Same model published on other platforms. May have additional downloads or version variants.