Light Color1.0

The training datasets used for Light Color 1.0 and Light Color 2 are the same. What differs are their total training steps, as well as the values of network_dim and network_alpha. These differences have a significant impact on the results. You may have noticed that Light Color 1.0 is a full seven times smaller than Light Color 2. This is because I'm on a tight budget and my home computer doesn't have a large amount of memory. So when training the models, I hoped to achieve the best results with the least memory consumption. Due to the huge gaps in parameters such as the number of training steps, their effects are completely different when changing some detail - enhancing tags. When I used them, I added these tags to enhance the details of the images. You can give it a try.

masterpiece,best quality,Subtle Elegance,FluidLines,ElegantBrushwork,Impressionism,Masterpiece,Calm Design,concept art,lightly painted,gentle brush strokes,light and airy

If it's a full - body image, or if there are issues with the face or hands, it is recommended to use ADetailer to carry out the restoration of the face and hands.

The usage weight of LoRA：0.6~1

Sampler：Euler,Euler a

Steps：20~35

CFG scale：2~7

Monochrome triggers：

line style,monochrome,greyscale,simple background,sketch,

Color triggers：

colored sketch,traditional media,sketch,line style,simple background,

Minimalism

The usage weight of LoRA：0.4~1

Sampler：Euler,Euler a

Steps：20~35

CFG scale：1~7

Trigger words:

Minimalism,SimpleDesign,LineArt,simple background,

Light Color2

The usage weight of LoRA：0.6~1

Sampler：Euler,Euler a

Steps：10~35

CFG scale：2~7

Color Trigger words:

colored sketch,sketch,traditional media,

Monochrome Line Art Trigger words:

line style,monochrome,greyscale,sketch,traditional media

Background suggestions:

white background,simple background,sketch style,whimsical,soft lines,hand-drawn

DigitalArt

The usage weight of LoRA：0.4~1

Sampler：Euler

Steps：20~35

CFG scale：4~7

Trigger words:

DigitalArt,RoughTexture,BrushStroke

Prompt：

(best quality:1.1),(masterpiece:1.2),(detailed:1.1),(highres,  best quality:1.2),(ultra HD quality details),extremely delicate and beautiful,beautiful_detailed_eyes,highly detailed,crystalstexture skin,official art,

Negative prompt：

nsfw,worst quality,old,early,low quality,lowres,signature,username,logo,(bad hands:1.7),(mutated hands:1.6),mammal,anthro,furry,ambiguous form,feral,semi-anthro,poorly drawn face,disfigured,ugly,(missing fingers:1.4),(malformed hands:1.5),(poorly drawn hands:1.4),too many fingers,fused fingers,stare eyes,wide-eyed,skin blemishes,acnes,

Watercolor

Hs Art Anime-Illustrious

The usage weight of LoRA：0.3~0.9

Sampler：Euler

Steps：10~35

CFG scale：2~7

Trigger words:

watercolor

Auxiliary trigger words:

ink,brush stroke,watercolor \(medium\),

Background suggestions:

watercolor paper texture,textured background,watercolor texture,rough paper,grainy texture,subtle texture,background texture,

Prompt：

masterpiece,best quality,newest,absurdres,highres,safe,beautiful_detailed_eyes,light_eyes,

Negative prompt：

stare eyes,wide-eyed,

nsfw,worst quality,old,early,low quality,lowres,signature,username,logo,(bad hands:1.7),(mutated hands:1.6),mammal,anthro,furry,ambiguous form,feral,semi-anthro,poorly drawn face,disfigured,ugly,(missing fingers:1.4),(malformed hands:1.5),(poorly drawn hands:1.4),too many fingers,fused fingers,

Flat Design

The usage weight of LoRA：0.6~1

Sampler：Euler,Euler a

Steps：10~35

CFG scale：2~7

Trigger words:

Flat Design

Auxiliary trigger words:

vector art,simplistic style,minimalistic

Background suggestions:

simple background,flat color

Facial details:

no facial features

Silhouette Art

The usage weight of LoRA：0.6~1.4

Sampler：Euler

Steps：10~20

CFG scale：2~7

Background suggestion: Choose a solid color background

Trigger word：SilhouetteArt,monochrome,greyscale,high contrast,spot color

Line Art

This model is mainly designed for the training of artistic line styles. You can choose either a black or a white background. In this way, the resulting images will have a cleaner look and the artistic effect will be really excellent. When it comes to the overall coloring of the picture, you can opt for either a monochromatic or a colorful artistic effect.

If you want to achieve a monochromatic effect, you can enter tags such as "greyscale" and "monochrome". However, if you hope to get a colorful line effect, please don't enter tags related to monochromatic styles.

Regarding image generation, if you're not familiar with the models, I really don't recommend that you use more than three LoRA models created by different authors. This is because these LoRA models may not necessarily be compatible with each other, and the resulting image quality can be truly unappealing. I really can't understand the thinking of some people. Do they really think that the more LoRA models they use, the better the effect will be? The prerequisite is that you should know what function each LoRA model serves!

这个模型主要是针对艺术线条风格的训练，你可以选择背景是黑色，或者白色，这样出的图效果会更加干净艺术效果非常好。画面整体上色你可以选择单色或者是彩色的艺术效果。 如果你希望呈现单色的效果，可以输入greyscale, monochrome这些tag，如果你希望呈现彩色的线条效果，那么请不要输入单色等tag。

对于出图方面，如果您不了解模型，我实在不建议您使用超过3个以上且出自不同作者的模型LoRA，因为这些LoRA不一定能兼容在一起，出的图的效果真的难以直视，我实在不明白有些人的想法，他们是否以为LoRA用的越多，效果就会越好呢？前提是你知道每个LoRA起到的作用是什么吧！

The usage weight of LoRA：0.2~0.9

Sampler：Euler

Steps：10~20

CFG scale：2~5

For the basic model, I recommend that you use: Hs Art Anime-Illustrious.

Coloring style

line art,sketch,draft,abstract lines,minimalist,dynamic lines,Hand-drawn style,

Black background

line art,monochrome,greyscale,simple background,black background,sketch,draft,abstract lines,minimalist,dynamic lines,Hand-drawn style,

White background

line art, greyscale, monochrome, sketch, white background, draft, abstract lines, minimalist, dynamic lines, Hand-drawn style