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
对比1.2.1版本优化了构图和提示词遵从度，应该是1.2.X版本的最后一次更新
Compared with version 1.2.1, the composition and prompt word compliance have been optimized. This should be the last update of version 1.2.X
FAQ
Comments (49)
sometimes I get a really old lady result like 60s something with sunken cheek(especially this one) and wrinkles. What is the recommended negative tag to avoid this ?
Old lady, old woman, wrinkles or something else ?
Also is there an age tag like 30s, 40s woman ?
If facial restoration is used, do not add words like "face" or "face focus".
The age tag is not included because there is no anime dataset containing the age tag
EKLL ahh, so If I use afterdetailer I need to avoid promp word any "face" in general right ? thank you.
clayr yes
Sorry, after ten days of testing, I found that the wrinkle issue is related to an unreleased lora added when making the model
Can you add it to tensorart?
If there is a chance
EKLL Please
EKLL yeah please do 🙏
robertomanfinfla I understand everyone's needs, but there is a reason why I don't maintain the tensorart version.
First of all, when everyone uses my model on tensorart, the rebate points cannot be used for the cloud computing function of tensorart. Because if you don't have a tensorart membership, you can only use a dataset of up to 100 images during training, which is quite restrictive and basically useless.
Secondly, the network traffic required for uploading the model is limited, which means it is necessary to choose to maintain only a few websites.
Please understand. I will consider your needs as appropriate.
haiqaljihad583 I understand everyone's needs, but there is a reason why I don't maintain the tensorart version.
First of all, when everyone uses my model on tensorart, the rebate points cannot be used for the cloud computing function of tensorart. Because if you don't have a tensorart membership, you can only use a dataset of up to 100 images during training, which is quite restrictive and basically useless.
Secondly, the network traffic required for uploading the model is limited, which means it is necessary to choose to maintain only a few websites.
Please understand. I will consider your needs as appropriate.
EKLL Ok
I hope the author can improve the compliance and accuracy of the prompt words in the next version. Currently, Wai and hass are out, which is my first choice. Keep it up. This checkpoint is very good. For those who like mature women, this is simply heaven.
thank you
EKLL Hi, excuse me, I would like to ask a question that I often encounter. When generating NSFW images with sex scenes or insertions like dildos, how do I control the subject and the inserted object? Sometimes the images I generate often have the object reversed.
wandong1 Describe it in short natural language sentences
I need this checkpoint urgently in tensorart. Civitai won't let me use it. :/
The next version may synchronize tensorart separately
花费了10天测试了超过100种新的融合比例，这个“新版本”在各方面很好，但无论如何都无法调整年龄。所以10天被浪费了，新版本仍旧遥遥无期。
I spent 10 days testing over 100 new merge ratios. This "new version" is great in all aspects, but no matter what, the age cannot be adjusted. So ten days were wasted and the new version is still a long way off.
大佬能放出来试试嘛 （˶･֊･˶）
Rienie 感觉没有发布的意义，和其他模型一样只能生成萝莉，最多到大学生？大概吧
EKLL 那樣就不是Mature Ritual了。。。。。。不過之前在X上我記得你不是還做了一個LifeStage的V-pred版本嗎？什麽時候會發佈那個模型呢。。。。。。
LOL2024 所以这个版本已经被删除了，我自己都没留档。LifeStage的V-pred版本微调后已经作为Mature Ritual V-pred发布了。LifeStage的V-pred版本项目实际上已经不存在了，我也在考虑LifeStage模型的存在必要性，下载量实在过低了，又有上手难度，感觉不太友好，准备好Mature Ritual新的基准版本之后，应该会着力与GrownRuler v0.4的数据集制作（可能会面向大家募集图像数据）
Version 2 is still the best of all. Maybe you could create a new version with the latest characters and add more male characters so the MILFs can have fun!
thank you
征集一些大家想要测试的提示词
Collect some prompt words that everyone wants to test
试试sex,sex from behind,standing sex,bent over,standing,grabbing another's breast,grabbing from behind？感觉后入抓胸的效果特别差
@ddmf098 sex from behind,standing,bent over,grabbing breast,
@EKLL 1boy and 1girl,uncensored, ,from behind,
missionary, sex, leg lock, soles, tiptoes,hug,
MTSPJIN,,brown eyes,orange hair,hair bun,
huge breasts,hetero,long legs,skinny,narrow waist, thin thighs,nipples,
see-through babydoll, panties around one leg,
blush,heavy breathing,moaning,saliva,drooling,shy,embarrass,eye contact,
oreki houtarou,(faceless male),(shota:0.8)
on back, on bed,
large penis,cumdrip,pussy juice,testicles,
indoors, room,night, warm lighting,
trembling,motion lines,spasm,
用的v1.3，感觉上面的提示词很容易生成没有手，或者肢体不协调的男性
@ddmf098 by MTSPJIN,
1boy,oreki houtarou,1girl,uncensored,from behind,
missionary sex, leg lock, soles, tiptoes,hug,
brown eyes,orange hair,hair bun,
huge breasts,hetero,long legs,skinny,narrow waist, thin thighs,nipples,
see-through babydoll, panties around one leg,
blush,heavy breathing,moaning,saliva,drooling,shy,embarrass,eye contact,
(faceless male),(shota:0.8)
on back, on bed,
large penis,cumdrip,pussy juice,testicles,
indoors, room,night, warm lighting,
trembling,motion lines,spasm,
我测试了随机种子数*3，图片发到下面了，您可以看看
@EKLL 感觉最右边图里男人的手不太协调？
@ddmf098 三分之一的概率算很差吗。。。
@EKLL 其实中间和最右边的图肢体协调性在我看来都有点奇怪。可能是我的要求太高了。还是感谢回复
@ddmf098 还可以调整画布大小，方法有很多。。。测试图我都默认832*1216的
感觉提示词的准确度还是不够好，肢体的协调性还是差了点，尤其是生成两个人的时候。不知道是不是我提示词的问题
是不是添加了质量词？顺便用来测试的提示词是什么
@EKLL 两人正常体位的时候，男人的手有时候没有，或者手的位置不协调
@ddmf098 应该完全不使用含义冲突/重复的词，比如，扶着、腰，应该直接写成扶着腰，或者holding、weapon、holding weapon应该直接写成holding weapon
@EKLL 我感觉ai其实分不太清男女？我指定男女做某些动作，结果ai会搞不清
@EKLL 用后入位的话，grabbing another's breast经常会出现断手的情况
@ddmf098 不要用非常规的图片尺寸，成熟的合并模型出现肢体错乱的情况一般会是lora冲突和图片尺寸还有爆显存三种情况造成的
@zxj50751055729 我一般用1024*1024或者832*1216，这样算非常规的图片尺寸吗？
@ddmf098 尺寸没问题
我尝试将模型与wai模型融合，但是报错了。chatgpt发现的错误是TypeError: argument 'metadata': 'dict' object cannot be converted to 'PyString'意思是：程序在保存模型的时候，给 metadata 参数传了一个“字典 (dict)”类型的东西，但它只接受“字符串 (string)”类型，所以报错了。我想进行模型融合，大佬能告诉我解决方法吗。