WAI-REAL_CN

CHECKPOINT
Original


Updated:

107K

关于例图, 我使用了1024x1360的大小,都是直接输出没有经过AD跟高修.

For the example images

V13, I used 1024x1360,

generated directly without AD or hires fix.

V13

v13主要是进行了色调跟数据调整,对脸部基本没有太大调整.

Version 13 mainly focuses on color tone and data adjustments, with minimal changes to the face.

把cfg稳定在了7,方便后面版本的调整.

The CFG has been stabilized at 7 to facilitate adjustments in future versions.

v13 Recommended Setting

  • Steps: 30

  • CFG scale: 7

  • Sampler: Euler a

Positive Prompt

score_9, score_8_up, score_7_up,source_real,real skin, BREAK

Negative Prompt

worst quality,low quality, source_furry, source_pony, source_cartoon,2d, ,bad anatomy, bad hands,blurry,3d,

v12跟12step同时发布

关于这2个版本,我对精度进行了提升,可以不需要太依赖于AD进行脸部修复.然后还进行了一些背景的优化.

For these two versions, I have improved the precision, so there's no need to rely too much on AD for face restoration. Additionally, I have made some optimizations to the background.

关于例图, V12版本我使用了1024x1360的大小 12步使用了832x1216. 都是直接输出没有经过任何其他插件的优化.

For the example images

V12, I used 1024x1360,

12 steps, I used 832x1216.

Both were generated directly without any optimization from other plugins.

v12 Recommended Setting

  • Steps: 30

  • CFG scale: 5~7

  • Sampler: Euler a

12 steps Recommended Setting

  • Steps: 12

  • CFG scale: 1

  • Sampler: Euler a

two versions same Prompt

Positive Prompt

score_9, score_8_up, score_7_up,source_real,real skin, BREAK

Negative Prompt

worst quality,low quality, worst detail,low detail, source_furry, source_pony, source_cartoon,2d, ,bad anatomy, bad hands, text, error, missing fingers, (extra digit:1.4), fewer digits, cropped, bad artist, outlines,jpeg

_______________________________________________________________________

V11

About V11

对v10版本整体权重调整以及进行了一点摄影方面得训练跟调整.

在光影,纹理,镜头表现方面有一定得优化.

The user has made overall weight adjustments to the v10 version and has also conducted some training and adjustments related to photography.

The user has made optimizations in light and shadow, texture, and lens performance.

V11

Recommended Setting

  • Steps: 30

  • CFG scale: 5~7

  • Sampler: Euler a

  • ADetailer face_yolov8s.pt use (ADetailer sampler: DPM++ 2M, ADetailer scheduler: Karras )

Positive Prompt

score_9, score_8_up, score_7_up,source_real,real skin, BREAK

Negative Prompt

worst quality,low quality, worst detail,low detail, source_furry, source_pony, source_cartoon,2d, ,bad anatomy, bad hands, text, error, missing fingers, (extra digit:1.4), fewer digits, cropped, bad artist, outlines,jpeg artifacts, signature, watermark, username, blurry,3d,

__________________________________________________________________________________________________

关于v10下面会有很长一段的文字说明,如果你有时间的话请务必读一下,我会很感谢你.

如果你不想知道这些.请往下拉直接观看使用说明.

Regarding V10, there will be a long explanation below. If you have the time, please take a moment to read it—I would greatly appreciate it.

If you're not interested in this, feel free to scroll down and go straight to the usage guide.

________________________________________________________________________________________________

关于v10,最近和几个朋友交流的时候发现一件事,我好像过于追求真实的皮肤真实的脸部.回想起来我做这个系列模型的初衷是为了好看的亚洲图片,而不是为了真实而真实.所以我以后的模型升级方向是为了更加好看的图片,所以我会尝试各种方法,如果后续模型的写实程度低于以前的请见谅.

Regarding V10, I recently noticed something during conversations with a few friends: I seem to have been overly focused on achieving realistic skin and facial features. Reflecting on it, the original goal of this model series was to create beautiful Asian images, not to pursue realism for the sake of realism. Therefore, my future model upgrades will focus on creating more aesthetically pleasing images. I will be experimenting with various methods, so if the realism of future models is lower than before, please understand.

__________________________________________________________________________________________________

V10改动

1.我使用V9进行了一部分素材的追加训练得到的新模型,然后跟我以前训练的素材模型融合以及lora的调整,使得V10的亚洲人的权重进一步增加,现在的结果就是V10生成的脸部都会偏向亚洲特征,其他人种的特征会很不明显.

2.其次我对写实的权重也放低了几个等级,现在默认的写实程度是要低于v9的,所以必须通过增加写实tag跟调整cfg来得到更高的写实效果,至于我为什么要这么做呢,因为有些时候不是写实的图片就好看,我更喜欢那种带有写实风格的但是人物其实并不会太写实的那种效果.所以我为了达到效果,所以做了那么麻烦的改动,使用的时候虽然增加了劳动,但是会得到更多有趣的结果.

V10 Changes

1. I created the new V10 model by conducting additional training on some materials using V9. I then merged it with previously trained material models and adjusted the LoRA settings. As a result, the weight of Asian features in V10 has been further increased, so the faces generated by V10 will predominantly exhibit Asian characteristics, while features of other ethnicities will be much less noticeable.

2.Additionally, I have lowered the realism weight by several levels. Now, the default realism in V10 is lower than in V9, so achieving a higher level of realism will require adding realism tags and adjusting the CFG scale. The reason for this change is that realistic images aren't always the most visually appealing. I prefer a style that incorporates a touch of realism without making the characters overly realistic. To achieve this effect, I made these complex adjustments. While this might increase the effort during use, it will also yield more interesting results.

__________________________________________________________________________________________________

V10 简单使用说明

V10 Simple Usage Guide

1.对于只想简单生成图片的用户,请你务必按照我的推荐设置来使用,尤其是负面,请不要随意改动.CFG请只使用7.

1.For users who just want to generate images simply, please make sure to follow my recommended settings, especially for negative prompts. Please do not change them arbitrarily. CFG just use 7.

2. 对于上级用户,我的推荐是删除正面提示里面的"realistic,photorealistic, ",然后请从cfg3.5开始生成如果效果不满意再请增加cfg最大到7.如果结果还是不满意,在考虑增加写实tag.

2.For advanced users, my recommendation is to delete Positive Prompt"realistic,photorealistic, " and start generating with a CFG scale of 3.5. If the result is not satisfactory, please increase the CFG gradually, up to a maximum of 7.If the results are still unsatisfactory, consider adding realistic tag.

3.关于负面提示,请不要随意删除,这会影响生成图像效果.当然对于上级者来说请随意编辑,如果有更好的提示请务必给我留言.

3.Regarding negative prompts, please do not delete them arbitrarily, as this can affect the image generation quality. However, advanced users are welcome to edit them freely. If you have better prompts, please make sure to leave me a message.

V10

Recommended Setting

  • Steps: 30

  • CFG scale: 3.5~7

  • Sampler: Euler a

  • ADetailer face_yolov8n/s.pt use (ADetailer sampler: DPM++ 2M, ADetailer scheduler: Karras )

Positive Prompt

score_9, score_8_up, score_7_up,source_real,realistic,photorealistic, BREAK

Negative Prompt

worst quality,low quality, worst detail,low detail, source_furry, source_pony, source_cartoon,2d, ,bad anatomy, bad hands, text, error, missing fingers, (extra digit:1.4), fewer digits, cropped, bad artist, outlines,jpeg artifacts, signature, watermark, username, blurry,3d,

__________________________________________________________________________________________________

V2toV9

Recommended Setting

  • Steps: 30-50

  • CFG scale: 7

  • Sampler: Euler a

  • Sampler: Euler a ADetailer face_yolov8n/s.pt use more better

Positive Prompt

score_9, score_8_up, score_7_up,

Negative Prompt

score_6,score_5,score_4, worst quality, low quality, bad anatomy, bad hands, missing fingers, fewer digits, source_furry, source_pony, source_cartoon,3d, blurry,

________________________________________________________________________________________________

v8 发布

关于V8,首先我把v7的权重又往v6那边靠近了,亚洲占比更高,但是其他人种的数据还是比以前更高一点,如果要切换请把使用(国家+girl)的提示词增加权重.人种效果图我都在例图里面放了.

其他就是根据身边反馈的日常优化及其调整.关于灵活性方面,就我自己测试的来说应该是足够日常使用了,lora这块我还没仔细测试,有问题请跟我反应.

PS. 调整大模型是真的掉头发,本模型可能要很长一段时间不更新了.除非有非常大的bug.如碰到请跟我反馈.

v6->v8

  • 脸模参数调整

  • 增加了真实素材

  • 调整了灵活度

  • 更自然的打光

v6->v7

  • 调整的光影

  • 进一步加强了一点灵活性

  • 增加了一些背景素材

v5->v6

  • 调整了灵活度

  • 修复了一些问题

v4->v5更改

  • 调整了整体亮度

  • 增加了皮肤质感

  • 修复了一些问题

v3->v4 更改

  • 增加了一些背景素材

  • 继续增加了一些二次元数据,不够写实请自行添加realistic等tag

v2->v3更改

  • 修复了一些问题

  • 增加了模型的泛用性

  • 添加了一些二次元的数据,如果不够写实请自行添加realistic等tag.

Same Recommended Setting With V2

__________________________________________________________________________________________________

v1-v2>更改点

  • 更改了底层能生成更多内容

  • 皮肤更接近黄种人

  • 背景优化

  • 光照优化

  • 修复了一些问题

_________________________________________________________________________________________________

V2由于底层的改变参数也有所改变

v2 Recommended Setting

  • Steps: 30-50

  • CFG scale: 7

  • Sampler: Euler a

  • Sampler: Euler a ADetailer face_yolov8n/s.pt use more better

Positive Prompt

score_9, score_8_up, score_7_up,

Negative Prompt

score_6,score_5,score_4, worst quality, low quality, bad anatomy, bad hands, missing fingers, fewer digits, source_furry, source_pony, source_cartoon,3d, blurry,

________________________________________________________________________________________________

第一版 欢迎多返图跟提意见 后面会根据反应来改进(大概 如果有空的话)

Recommended Setting

  • Steps: 30

  • CFG scale: 3.5-7

  • Sampler: Euler a or DPM++ 2M Karras

  • ADetailer face_yolov8n/s.pt use more better

Positive Prompt

score_9, score_8_up, score_7_up,

Negative Prompt

score_6,score_5,score_4,bad anatomy, bad hands, missing fingers, fewer digits, source_furry, source_pony, source_cartoon,3d, blurry,

Version Detail

Pony

Project Permissions

    Use Permissions

  • Use in TENSOR Online

  • As a online training base model on TENSOR

  • Use without crediting me

  • Share merges of this model

  • Use different permissions on merges

    Commercial Use

  • Sell generated contents

  • Use on generation services

  • Sell this model or merges

Comments

Related Posts

Describe the image you want to generate, then press Enter to send.