我们需要真正理解女性年龄的模型而不是只能生成萝莉！
We need a model that truly understands female age, not one that only generates "loli"!
所以这是一个旨在生成成熟女性并且试图理解多样的女性年龄，而不是只能生成小孩、女孩、萝莉的模型。
Therefore, this is a model designed to generate mature female and attempts to understand the diverse ages of women, rather than only generating "children, girls, or loli."
我的推特:@ekll01
My Twitter: @ekll01
版本说明请查看About this version
Please see "About this version" for the version description.
Please refer to the workflow that I have disclosed.
目前支持的年龄tag
女性：lxlx、mesugaki、bishoujo、lady、mature lady
男性：shxtx、bishounen、gentleman
请您用以上tag代替1girl、1boy。
Currently Supported Age Tags
For females: lxlx, mesugaki, bishoujo, lady, mature lady
For males: shxtx, bishounen, gentleman
Please use these tags instead of 1girl or 1boy.
Recommended Settings（general）
CFG: 5
Steps: 40
Sampling: Euler a
Scheduler: simple
Resolution: 832*1216 above
希望喜欢这个模型的朋友能够量力为模型进行一下Bid投资。这不是请求，但是如果有人能够帮助模型进行Bid，我会十分感谢！
I hope that friends who like this model can consider making a "Bid" investment in the model according to their ability. This is not a request, but I would be very grateful if someone could help the model with a "Bid."
Description
This version is based on rouwei-v0.1 to remove the correction class of lora, a more pure version, convenient for people who want to train lora or fusion model use
这个版本是在rouwei-v0.1的基础上删除了修正类的lora，比较纯净的一个版本，方便想训练lora或者融合模型的人群使用
FAQ
Comments (58)
I'm not loving the new rouwei versions. I'll have to try out the latest one to see if there are improvements. Rouwei SEEMS like it could generate more interesting/dynamic results, but as I've used it all I've gotten is lower quality images compared to v0.4 in every regard. Maybe I'll just have to really change up my prompting structure to accommodate.
I've adjusted my prompting structure, and now my rouwei prompts come out MUCH better. Using 'BREAK' seems to really help rouwei models along and showcase its quality. Cheers!
@Terrordacturne This version is for the need to train lora or as a base mode to use, essentially rouwei version of the concise version
rouwei新版本，我如果正常生图，使用这两个版本哪个好，新版本跟上一个版本有啥区别
basic版本删除了一些修正类的内容，更加原始，适合需要训练lora或者作为底模进行融合的人，正常生图上个版本成图率更高，这个版本更原始
After testing all the versions, I'd rate them like this:
1. 0.3 and 1.0 EXP are the best overall, they are good at styles, composition and prompt adherence.
2. ONEPRO and TWO are the most stable, almost no problems with anatomy, work well with variety of negatives and prompt structures. However, they might give you strange results with specific tags.
3. All the rest are interesting experiments, but they are either heavily styled by default, or don't follow the prompt as good as other versions.
On the other hand, all this made me download Grown Ruler model, I want to test it too now. All the Mature Ritual models are still good compared to many other mixes.
You can try the ruowei version, I personally think it is not bad, Grown Ruler is currently abandoned update situation, the existing version may be a little old
Rouwei-basis-v0.1 is actually the Grown Ruler of the new version, but I put it under the Mature ritual
@EKLL I tired Rouwei (not the "basis" because it reads "for people who want to train lora or fusion model use", so I assumed it's not for usual generation), and it was ok, but something was off with generating retro-style images, and it was much worse at generating in Masamune Shirow style, for example (previous models are quite good at it).
I'll try Rouwei-basis, but I still want to try Grown Ruler because I've missed it back then.
@dobomex761604 The Artist tag of Masamune Shirow is shirow_masamune. Please use by shirow_masamune as the prompt word, which can be recognized in my test
@EKLL yes, I now this) My point is, previous Mature Ritual models are exceptionally good at this style - much better than other Illustious-based mixes, but Rouwei specifically is either too strong with this style, or just too weird (too anorexic, slightly worse anatomy). Even 0.4 is good with Shirow's style, I uploaded a picture with it back then (also, I have pictures with that style for earlier versions).
@dobomex761604 This may be the problem of ruowei, about ruowei and Illustrious is a trade-off, ruowei has a better character recognition, stronger conceptual understanding, the corresponding problem is also more, this may need the author of ruowei release a new version to solve, no version is absolutely perfect, suitable for their own is the best, I will also wait for ruowei I will wait for the author of ruowei to release a new version before updating. (I personally think that my Illustrious version is as good as I can make it, and updating it again won't bring any higher quality)
Sorry, I have been testing the ruiwei version for about one night, but there are some problems with it. In most cases, 1.0 EXP is better.
@EONICS I use Krita with plugin (which is essentially Comfy hidden under the UI), and all my images were generated with 1060 3gb. Always generate locally if you can.
@EKLL I'm currently testing Rouwei-basis-v0.1, and it's sooo different. It works well with styles, but, for example, DPO-like prompts don't really work with it. But if it's an experiment, it's definitely in the right direction. There's something in it, will post some pictures later.
@dobomex761604 According to my own personal feeling, ruowei is better in light and shadow, art style (non-artists tag) and Character quantity. In general, each has its own advantages and disadvantages. I am trying to merge ruowei and Illustrious
@dobomex761604 I've been trying for a few days but can't reproduce the style of this image from https://civitai.com/images/60067584
@EKLL That's probably because I used sgm_uniform scheduler. Civitai doesn't have separate schedulers in choice, and it doesn't mention post in the image itself https://civitai.com/posts/13431377
(and Krita plugin doesn't write metadata into images automatically, at least the old version I use)
@dobomex761604 webui or comfyui？
@EKLL Krita plugin uses Comfy UI as a backend.
Personally, two is the best for now.
I have switched between dozens of models so far, but ever since I came across this model, I haven’t used anything else. It’s truly amazing. I imagine creating models is a lot of hard work, but I hope you will consider creating a model with the same concept when the next-generation checkpoint emerges.
Thank you for your use and I will consider it
请问我要生成一个手伸进衣服里揉胸的提示词是什么，我用hand under clothes，grabbing another's breast，没有效果，只是隔着衣服抓
如果在danbooru上找不到对应的标签，且自然语言也不行的话，可能需要lora解决
This model is pure gold, the Illustriousv1.0EXP for being experimental is quite good, thanks!
Thank you for using my model!
大佬，我听说3月开始C站创作者能靠buzz打赏盈利了，还会继续更新下去嘛
As a milf lover, you model is easily my favorite. But i found that the eyes isn't train very well, if not to say very bad compare to other popular model. :(
This involves a lot of issues, please try to open facial repair
Can u release a model that doesn't have anime/asian faces? I'm trying to train western comic characters but the face gets influenced like asian/anime.
You may need a western comic lora. It's the easiest way.
The new version of Mature ritual could not satisfy me and I decided to postpone the release to recreate it.
Looking forward to testing anything new you will release. Personally I avoided your rouwei releases beacuse I am more on the Cartoon-Style and less on the Anime-Style,
Can anyone explain to me what all the hype around Noob-Models is about? I tried several checkpoints and never achieved anything as good as with IL checkpoints. Am I missing something crucial?
@houseconspiracy535 Thank you for your use
Will you bring GrownRuler back to the site? It's a very fine model.
@allamalla This model had errors in civitai, and I had no intention of updating it, so I deleted it.
There are also some conflicts in the positioning of "GrownRuler" and "Mature ritual" (" GrownRuler "is a" Mature ritual "that slightly adjusts age and does not add Greasy).
At present, "Mature ritual" and "JuicyArtRitual" will be updated, and it may be considered to merge "GrownRuler" into the version difference of "Mature ritual".
@allamalla Thank you for liking GrownRuler.
@EKLL I see, that's unfortunate as my local results in reforge were exceptional! With shiny/oily in negative, would v0.3 be the closest to the GrownRuler recipe?
@allamalla rouwei-basic-v0.1 is closer to the concept of "GrownRuler",
The civitai online generation feature is now live, you are welcome to try it out!
By the way, civitai has launched the resource recommendation bidding function, I hope those who have spare buzz can support me more, I really need more exposure and buzz!
This is amazing, thank you!
Which of your styles do you like the most? I didn't have desirable results with rouwei and exp to be honest.
I still use OnePro a lot.
@FondantViper Thanks for your use, you can try Illustriousv1.0EXP
Hello, I can't understand the version scheme. which is your current version as of FEBRUARY 2025? Thank you very much.
A new version will be released tomorrow, so you can try the new version directly
how do you train your models to make the output look so great? are they trained on specific artists?
@EKLL the new model looks great too!
@mynamemt1414 thank you
@EKLL is there a way to replicate looks of v0.2 (older mature) in the newer versions?
@mynamemt1414 There are ways to do this, but in my experiments, over-adding a V0.2-like look and feel reduces compatibility and diversity
大佬，生成过程中图片是正常的，但是生成出来变成马赛克颜色很奇怪是什么情况，求解答
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.