The color and left-right combination of the heterochromia eyes can be fixed.
Total of 12 combinations of four colors (blue, green, yellow, and red), each with left and right sides, are available.
There are a few quirks to using this LoRA. Please refer to the "Usage" section.
Reproduces mouths pronouncing Japanese 5 basic vowels, "A" (Ah; /a/) , "I" (Ee; /i/) , "U" (Woo; /ɯ/) , "E" (Eh; /e/) , "O" (Oh; /o/) .
It can be applied to a variety of other applications depending on its shape.
Reproduces sensual (voluptuous) face with half-closed (and a bit wet) eyes and inverted-v-shaped eyeblows. Also suitable for NSFW scenes.
4 types are available.
This LoRA reproduces smirking eyes, V-shaped eyebrows, widely open smug mouth with single fang. Girls with this face is called "Mesugaki" in Japanese 2D illustration meme.
Will make better shapes than prompt-only.
This LoRA reproduces widely open smug mouth with single fang. Will make better shapes than prompt-only.
Single fang(double tooth) is called "Yaeba(八重歯)" in Japanese 2D illustrations, represents innocence or childhood.
This LoRA reproduces eyebrows with the ends sloping downward (inverted V-shaped), which are often used to convey expressions such as troubled or sensual looks.
While similar eyebrows can be reproduced by specifying some expressions in prompts, using this LoRA allows you to avoid fixing expression prompts, making it easier to combine with other expressions.
Reproduces V-shaped eyebrows to express smug / proudly face (called "Doyagao" - Japanese anime slung).
Usually, V-shaped eyebrows can be reproduced by using V-shaped eyebrows prompt, but it sometimes makes very thick eyebrows. This LoRA does not reproduce thick one.
With most checkpoints, "o"-shaped (circular) mouth can be reproduced with prompt (:0), but its size may be larger or smaller than expected.
With this LoRA, mouth size can be fixed to large size or small size.
When using Stable Diffusion, basically front side of the character is lit up, making it difficult to reproduce a dark state.
With this LoRA, it is easier to reproduce a dark state with almost no light on the front of the character.
In addition, a LoRA is also available that makes it easier to reproduce the "glowing eyes" often used for characters in the dark as a dramatic effect.
To reproduce closed eyes, usually closed eyes prompt is used. But it may not certainly reproduce closed shape, sometimes get wink or half closed eyes.
This LoRA helps to reproduce smiling faces with better shaped closed eyes. The eyes will have a stronger upward curve than the normal closed eyes prompt.
To reproduce closed eyes, usually closed eyes prompt is used. But it may not certainly reproduce closed shape, sometimes get wink or half closed eyes.
This LoRA helps to reproduce thoughtful look with better shaped closed eyes. The eyes will have a stronger downward curve than the normal closed eyes prompt.
Reproduces a face strongly turned red with embarrassment.
Three types are available: one with a red line on the face, and two types with no red line but only a red tint (full face/cheeks only).
