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

V3.0 Update

来来回回测试了三个月,总算是有阶段性的成果了

改用LyCORIS进行训练,以保留更多的服饰细节,模型需要使用a1111-sd-webui-lycoris插件加载

往数据集里新添加了一些素材,优化打标方式,现在可以将几部分的特征比较好地分离出来,主tag还是nahida \(genshin impact\),主要负责头部除头饰和瞳孔以外的特征生成,四叶印瞳孔的tag是cross-shaped pupils ,但目前效果不太理想。想使用原版服装的话可以根据需要添加下列词条:in-game dress, hair ornament, bloomers, stirrup legwear, bracelet, cape,想换装就参考图例用其他tag

推荐unet权重0.4~1.0,低权重下可能需要手动添加耳朵、发型tagpointy ears, multicolored hair, gradient hair, white hair, green hair来维持人物特征

如果感觉画面较脏,推荐使用ClearVAE改善画质

Translate with DeepL:

After three months of testing back and forth, we finally have a milestone result.

Change to use LyCORIS for training to retain more details of clothing, this version of model needs to be loaded using a1111-sd-webui-lycoris plugin.

To the data set to add some new material, optimize the tagging method, now can be better separated from the features of several parts, the main tag is still nahida \(genshin impact\), mainly responsible for the head in addition to headdress and pupils other than the generation of features, four-leaf print pupils tag is cross-shaped pupils, but the current effect is not ideal. If you want to use the original clothing, you can add the following tags as needed: in-game dress, hair ornament, bloomers, stirrup legwear, bracelet, cape, want to change the reference legend with other tag.

Recommended unet weight 0.4~1.0, low weight may need to manually add ears, hairstyle tags pointy ears, multicolored hair, gradient hair, white hair, green hair to maintain character features.

If you feel that the picture is a little dirty, we recommend using ClearVAE to improve the quality.

V2.0 Update

I learned a lot in this month, and my graphic card also suffered a lot :P, till now I can publish a better version for this model

For this version, recommend weight is 0.45~0.7, only one trigger word nahida \(genshin impact\) is needed, but has better restoration of characteristic

My suggestion is use lower weight for model, and higher weight for trigger word, so you can easily change her suit

All known problems in previous version are solved

The next version will be a LyCORIS model, please stay tuned!

Any suggestions are welcome!

V1.0 Release

Genshin Impact Nahida LoRA, training with base model final-prune[925997e9].

Recommend weight 0.45~0.8. around 0.7 is a sweet spot for customize and keep characteristic.

Models used for examples:

Use tags braid, sidelocks, side ponytail, multicolored hair, white hair, green hair, gradient hair to fix hair style.

Use tag bracelet to generate a in-game style bracelet (working better in higher weight).

Use tag white bloomer to add in-game style bloomer for charater.

Known problem:

  • Lower weight may cause a little incorrect on the green part of the hair color.

  • Sometimes a unexpected crown would appear on Nahida's head (mostly in higher weight).

Any suggestions are welcome!

Images Generated With This Model

Loading...

Similar Models

Keqing | Genshin Impact
V1
Keqing | Genshin Impact
(4.25)
0
8 Ratings
JUJUTSU SORCERERS | Depository of Jujutsu Kaisen Characters
V1
JUJUTSU SORCERERS | Depository of Jujutsu Kaisen Characters
(5.00)
0
3 Ratings
Nahida from Genshin Impact
Nahida from Genshin Impact
(5.00)
0
2 Ratings
Nazuna Nanakusa (Call of the Night) LoRA
V1
Nazuna Nanakusa (Call of the Night) LoRA
(5.00)
0
1 Ratings
Dottore | Genshin Impact
V1
Dottore | Genshin Impact
(5.00)
0
3 Ratings
Nahida v1.0
V1
Nahida v1.0
(5.00)
0
1 Ratings
FUTURE GADGET LAB | Depository of Steins;Gate Characters
V1
FUTURE GADGET LAB | Depository of Steins;Gate Characters
(5.00)
0
1 Ratings
Nazuna Nanakusa (Call of the Night) LoRA
V1
Nazuna Nanakusa (Call of the Night) LoRA
(5.00)
0
1 Ratings
JUJUTSU SORCERERS | Depository of Jujutsu Kaisen Characters
V1
JUJUTSU SORCERERS | Depository of Jujutsu Kaisen Characters
(0.00)
0
0 Ratings
JUJUTSU SORCERERS | Depository of Jujutsu Kaisen Characters
V1
JUJUTSU SORCERERS | Depository of Jujutsu Kaisen Characters
(0.00)
0
0 Ratings

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.