Update information
V3.0 updated. Fine-tuned LoRA to improve the effects of generating characters with complex body limbs and backgrounds. The overall styling is more toward manga style rather than simple lineart. A lower CFG scale can make lines thinner.
V3.0更新。对LoRA进行了微调，改善生成人物肢体和复杂背景的效果。整体更趋向漫画风而非纯线稿。低CFG scale可以让线条变细。
Recommend settings
Model: Anything V4.5
VAE: Orangemix (the same with NAI)
LoRA Strength: 1
Sampler: DPM++ 2M Karras
Sampling steps: 20
CFG: 7
Negative embedding: EasyNegative、badhandv4
Highres fix is also recommended.
Note: all the LoRA names used in sample images are my local name, you need to change them to your saved LoRA filename!
基模：Anything V4.5
VAE：Orangemix（同NAI）
LoRA强度：1
采样器：DPM++ 2M Karras
采样步数：20
CFG：7
负面embedding：EasyNegative、badhandv4
推荐使用高清修复。
注意：样图中的LoRA名称是我本地文件名，要改成你保存的文件名！
Difference
Here's a comparision of V2.0 and V3.0 (locally called V3 and V4 on my machine - V2 on my machine is not published):
V2.0和V3.0的对比（在我的本地机器上命名为V3和V4——本地的V2我并未发布）：
Comparision of different CFG scale:
不同CFG scale对比：
Description
Lineart style LoRA trained with offset noise, use monochrome or lineart to trigger, I highly recommend removing any color words from your prompt, like:
purple eyesblueflowerreddress
线稿风LoRA，使用时请添加monochrome 或lineart 开启。推荐将prompt中与颜色相关的单词去掉。
The demo images are generated w/ Anything V4.5 (orangemix VAE).
如何利用ControlNet提取线稿（using ContorlNet to extract linearts from illustrations）：https://www.bilibili.com/video/BV1Nv4y1a7ts
FAQ
Comments (22)
图生图跑不出来
用controlnet，不要用图生图
跑不出来，错误了
太棒了，可以用来生成controlnet用的图～
Great model with unlimited gameplay, bro. Well done！
Pretty neat, if you use the scribble control net with these results, you'll have a great reference image for later, that doesn't necessarily have to be anime.
I'm not sure which is better, canny or scribble. But anyway, neat.
I love this. The creators of controlnet supposedly had something like this in the works but never released it due to worry of repercussions or something like that. Glad someone said "screw that" and did it themselves.
I'd love to make a model on some line art of some of my own drawings (completely different style).
Would you mind sharing some insights as to how I can go about training? Do I treat it as a style when training?
Yeah I think so
一直想要这种线条的风格，太棒了！
God almighty if you could get western style illustration down like this... everyone would lose their minds.
很强力，基本画不出颜色
都是，text2img，有没有,img2img，把彩色转线条的安全
用ControlNet就行了
v3模型在哪里下载
怎麼碼都有色塊，不能只出線稿，請問哪裡出了問題呀...
去掉带颜色的描述词
Is there any model that already does the coloring? Or any idea how I can do this?
Hey, can you let me know what the base model used for this is?
I LOVE IT!