2.5D Chinese style healing concept.

0
0
0
0
CGI styleChinese StyleAnime StyleGirlStyle Boost
最近更新:首次發布:
CGI style,Chinese Style,Anime Style,LoRA,SD 1.5Image info
CGI style,Chinese Style,Anime Style,LoRA,SD 1.5Image info
CGI style,Chinese Style,Anime Style,LoRA,SD 1.5Image info
CGI style,Chinese Style,Anime Style,LoRA,SD 1.5Image info
CGI style,Chinese Style,Anime Style,LoRA,SD 1.5Image info

This is a painting style called Lora, which shows a good texture in the 2.5D and real human model format. At least, it doesn't make me feel that the colors are not mature or too mature. The example part is generated by Maigou DaDa's Lux or the realistic style of real people. The painting style is also significantly displayed in other large models, but I haven't tried too much. Specific details can be seen in the example image, I can't hold it anymore.

In short, different base models show different painting styles.

Version 1.0 has a weight of 0.6-1, you can try it yourself. This version is suitable for generating images with animals and can be used in conjunction with 1.5 or 1.6.

Version 1.5 has a weight that is optional, but not too low, preferably above 0.5.

Version 1.6 is a slight adjustment of version 1.5, with significant improvements in tone, painting style, and texture. The weight is optional, but it is recommended to be above 0.6 to show a more obvious effect.

I call this model a healing style national style, not a cartoon cute style.

Trigger word "hf_xy", for real people, it is recommended to use DPM++ SDE karras, but of course, this is not absolute. Depending on the combination of Lora or other models, other sampling methods can be used.

The effect of version 1.5 is good in the 2.5D large model, the real person author has not tried yet, everyone can try it. If you can't see it clearly, stick to 2.5D or anime style.

Versions 1.5 and 1.6 do not have animal elements, but enhance the character drawing. Because this is a style Lora, it can be used with other models. Of course, the original image is also good. It is recommended to add simple keywords, such as enhancing image quality, clarity, or describing characteristics of characters such as clothing color, hair color, and hairstyle.

Version 1.0 tends to be male-oriented in style, but most current models are female. However, based on my exhaustive testing, females also perform well. If changing gender does not change the result, please increase the weight, for example, (girl: 1.6) (boy: 1.6). If it still doesn't work, then please... continue to increase, don't ask me why, because the intensity is not enough.

Because it's a national style, I have also added zodiac elements to the model. This will be obvious when redrawing is done on a large scale when the corresponding tags are added, for example: dragon, pig, dog, tiger, snake....

Of course, you can also try adding "oversized_animal" when adding animal keywords. You can see that the animals have clearly become larger, while the characters become smaller.

In fact, you can also generate animals without characters.

As for other non-living entities, they have not been tested yet. If you are interested, you can try.

Please note that not all base models are suitable for this model. For example, if you insist on using a base model designed for girls to create an image of a boy, if you get a nice card, you can post it under the model for discussion and sharing...

The example images combine another Lora of mine, the Taoist robe, which has not yet been released here. If fellow Taoists need it, I will release it in liblib.

Version 1.5 has removed the weight restrictions for boys and girls, both genders are acceptable, and also removed the animal elements. If you want to have the 1.5 style, you can use it in conjunction with version 1.0.

The refinement of the model is not easy, I hope everyone will support it and return more images.

This model is for learning and communication purposes only and not for commercial use.

Enhanced texture, tone, and style based on version 1.5, can be directly output or used with other methods.

討論

最熱門
|
最新
發送
即將推出
下載
(0.00KB)
詳細資訊
類型
線上生成次數
0
下載次數
0
推薦參數
推薦搭配Checkpoint
推薦的重量
1
CFG
7
VAE

圖庫

最熱門
|
最新