Models

>

(female+male) Trafalgar law+chibi + 3suits +tattoo+sword DLC | one piece | 特拉法尔加 罗+Q版+三套衣服+文身+鬼泣剑 (海贼王)

(masterpiece:1.2), (best quality:1.2), hyper realistic,highestres,absurdres,extremely detailed,volumetic,dramatic lighting,
1boy,detailed eyes,perfect face,///,summer,beach, <lora:detailed_eye_V2-15:0.8:face3> , <lora:Trafalgar_law-10:0.8>,Trafalgar_law
(masterpiece:1.2), (best quality:1.2), hyper realistic,highestres,absurdres,extremely detailed,volumetic,dramatic lighting, 1boy,detailed eyes,perfect face,///,summer,beach, <lora:detailed_eye_V2-15:0.8:face3> , <lora:Trafalgar_law-10:0.8>,Trafalgar_law
masterpiece,best quality,highres,8k,extremely detailed,realistic, portrait,
1boy,without_hat,ice cream,tattoo,jacket, open clothes,earrings,///,summer, tree,sitting on bench,street, <lora:Trafalgar_law-10:1>,
masterpiece,best quality,highres,8k,extremely detailed,realistic, portrait, 1boy,without_hat,ice cream,tattoo,jacket, open clothes,earrings,///,summer, tree,sitting on bench,street, <lora:Trafalgar_law-10:1>,
masterpiece,best quality,highres,8k,extremely detailed,realistic, 
full body,
1boy,Trafalgar_law,tattoo,jacket, open clothes,earrings,sword, katana,///,summer,ice cream,tree,mountain <lora:Trafalgar_law-10:1>
masterpiece,best quality,highres,8k,extremely detailed,realistic, full body, 1boy,Trafalgar_law,tattoo,jacket, open clothes,earrings,sword, katana,///,summer,ice cream,tree,mountain <lora:Trafalgar_law-10:1>
Quick Generate
Select a style, write what you want to see and click Generate
What you want in the image.
Clear
Random
* This model is available for use and will generate similar character to the ones seen in the images.

Model Description by Creator

READ THIS BEFORE YOU DOWNLOAD

v5.5

-Eliminated the hat error in version 4 by PS refinement of some pictures

-Better compatibility and more flexibility with various models

-Re-examine each picture with labels

Because it took great effort and I don't have any income now (still in school), the content shown in the title and pictures are sold as virtual items. for more details see: https://ko-fi.com/7dragons/shop

Before this version are free.

Note: Because the kikoku sword is experimental, the second version is released for free, this is the thing that will get if you click the download button.

-通过对一些图片的PS改进,消除了第四版中的帽子错误

-对各种模型有更好的兼容性和更大的灵活性

-重新审视每张图片的标签

因为花了很大的精力,而且我现在没有任何收入,(还在上学) 所以标题和图片中显示的内容是作为虚拟物品出售的。更多细节见:https://ko-fi.com/7dragons/shop

在这个版本之前都是免费的。

注:因为剑是实验性的,所以第二个版本是免费发布的,这是如果你点击下载按钮就会得到的东西。

Please read the following to improve quality before use.

You can refer to my example picture to write prompt words

v4:

Each image was manually tagged and re-marked to remove images of lower quality.

The impact on the base model is greatly reduced.

Recommended weight 0.8~1, after the test, weight is 1 will not overfit, and basically restore the tattoo. Should be able to work if continue to increase, just try ; )

On tattoos:

  • shoulder tattoo,forearm tattoo,hand tattoo,finger tattoo,chest tattoo

They are:

  • a heart tattoo on the shoulder

  • a zigzag tattoo on the forearm

  • a cross tattoo on the back of the hand

  • a letter tattoo on the finger

  • and a large tattoo on the chest

Hint: In experiments, adding related tattoo words increases the chances of the corresponding tattoo appearing. Therefore, it will affect the movement and composition to some extent. Please add relevant words as you need.

  • "tattoo" should at least appear if you want to get the tattoo.

  • hand tattoo and finger tattoo are not recommended if there is a large number of hand deformities

  • If an unwanted close-up appears, put close up in the negative prompt word

  • The back tattoo can also be tried

prompts recommended:

  • Men's use of the tag word: 1 boy, law, earrings, hat, chest tattoo, hand tattoo, forearm tattoo, shoulder tattoo,

  • Women's use of the tag word: 1 girl, femlaw, earrings, hat, chest tattoo, hand tattoo, forearm tattoo, shoulder tattoo, cleavage, collarbone,

  • You can refer to my example picture to write prompt words

You can use the lora block weight addon to decrease the impact of style or color on the base model.

You can use this lora for inpainting , further improving the proposed redraw range of 0.55 for the tattoo section

Support me:

https://ko-fi.com/7dragons

for china:

https://afdian.net/a/7dragonsAI

Thank you very much! Your help makes me more motivated to create!

Thanks for reading ; )

————————————————————————————————————————

请在使用前阅读以下内容,提高图片质量:

你可以参考我的示例图片来写提示词

关于 V4:

  • 对每一张图片进行人工标签检查,并重新打标,去除了较低质量的图片。

  • 对基本模型的影响大大减少。

  • 推荐权重 0.8~1,经过试验,权重为1的时候不会过拟合,并且基本还原文身。应该可以继续增加,尽情尝试。

关于文身:

  • “shoulder tattoo,forearm tattoo,hand tattoo,finger tattoo,chest tattoo”

  • 分别是:肩膀上的心形文身,前臂上的锯齿状纹身,手背上的十字文身,手指上的字母纹身,以及胸前的大纹身。

  • 提示:经过试验,添加相关的文身词汇有助于提高相应文身出现的机会。所以会一定程度上影响动作和构图,请根据需要添加相关词汇

  • 不添加也可以,但提示词中至少应该出现tattoo以得到文身

  • 如果大量出现手部畸形,推荐不使用hand tattoo,以及finger tattoo

  • 如果出现不需要的局部特写,请将close up放到负面提示词

  • 附加:back tattoo也可以尝试

关于性别对应的提示词:

  • 男性罗使用的提示词:1boy,law,earrings,hat,chest tattoo,hand tattoo,forearm tattoo,shoulder tattoo,

  • 女性罗使用的提示词:1girl,femlaw,earrings,hat,chest tattoo,hand tattoo,forearm tattoo,shoulder tattoo,cleavage, collarbone

  • 仅供参考

可以使用lora block weight 插件改善本模型的风格或者颜色对底模的影响。

可以在图生图中使用此lora进行局部重绘,进一步改善文身部分建议的重绘幅度为0.55

支持我:

https://ko-fi.com/7dragons

爱发电:

https://afdian.net/a/7dragonsAI

非常感谢! 您的帮助让我更有动力创作!

感谢阅读 ; )

————————————————————————————————————————————————————————————————————————————————————————————————————————————————————————

v3

Except young, the other types do not need round hat or hat, because they are already included by default and will result in mixed hats.

建议权重0.7~0.8,0.8左右能较好复原文身,0.9会过拟合。 二次元模型基本都可以,真人模型要降低一点权重。

可以加入sword,katana来得到对应的武器。建议指定具体的衣服类型。

You can add sword and katana to get the corresponding weapon. It is recommended to specify a specific type of clothing.

触发词注释:

  • law: 戴帽子的男性罗,不用加hat这个词。

  • hatless:不戴帽子的男性罗。

  • femlaw:女性罗,基本戴帽子。不用加hat这个词。

  • young, round hat: 少年,也可以变成女生

其他的类型不用加 round hat or hat,因为默认已经包括,添加会导致出现混合帽子。没有想要的结果可以多抽几张

建议的 negative tags 添加:(Thigh tattoo:1.3),(leg tattoo),(Facial tattoo),(Facial mark),(neck tattoo:1.3),(stomach tattoo:1.2) ,waist tattoo,

题外话:

吁,累死我了。这个应该才算正式版吧,第一个就是随便做的。 在一天之中尝试了三次,并且colab刚刚还直接终止,图片没有来得及升级。 这个应该算第三版,训练集包括动漫截图以及同人图,第二版只有动漫截图,难以泛用,故而不发布,直接到V3。

网上的截图不好整,前几天为了找动漫截图重新看了一部分海贼王,截图截到手抽筋。头晕眼花,再加上网络环境不好,半卡不卡,断断续续,实在是考验我的耐心!整理了几个小时的标签,之前失败了还分析了一下原因。 所以,看在我这么可怜的份上,虽然没有物质回报,但是你可以点亮右上角的红心吗?你可以在下面多发几张图片让我欣赏一下吗?

Recommended weight 0.7~0.8, 0.8 can better restore tattoo, 0.9 will overfit.

The quadratic metamodel is basically OK, the real person model to reduce a little weight.

Trigger word note:

  • law: Men who wear hat, don't use the word hat.

  • hatless: A man who doesn't wear a hat.

  • femlaw: basically wearing a hat. Don't use the word hat.

  • young, round hat: the 10 year old law. Well, I guess he can turn into a girl, but I didn't get to try. These are the two trigger words, the first attempt.

The other types do not need round hat or hat, because they are already included by default and will result in mixed hats.

STRONGLY Suggested negative prompts to add: (Thigh tattoo:1.3),(leg tattoo),(Facial tattoo),(Facial mark),(neck tattoo:1.3),(stomach tattoo:1.2), waist tattoo,

Oh, I'm so tired. I think this is the official version. The first one was just random.

Tried three times in one day, and colab just stopped directly, the picture did not have time to upgrade.

This should be the third edition, the training set includes animation screenshots and the same figure, the second edition only animation screenshots, difficult to use, so not released, directly to V3.

Online screenshots are not good, a few days ago in order to find animation screenshots to see a part of One Piece, screenshot hand cramp. Giddy, coupled with the network environment is not good, intermittent, it is a test of my patience! I've been working with the tags for hours, and I've been trying to figure out why they failed.

So, for my pity's sake, and with no material reward, could you light the red heart in the top right corner? Can you add more pictures below for me to enjoy?

Images Generated With This Model

Loading...

Train Models

Train AI Models On Your Own Pictures

You can use your own pictures to train AI models. Reimagine yourself in new ways with our AI-powered image generator.
Create pictures of yourself around the world, in your dream outfit, as a video game character, or in countless art styles. The possibilities are endless.
Train Your Own Model

Generate AI Images. Simple and Easy

AIEasyPic is the #1 AI image generator. Generate images with simple prompts, swap faces, train models to reimagine yourself, and more.