有事加群：188546326
各风格图片参考下面链接
style Pictures refer to the link below
X/Y/Z
这是一些常用画风模型，由不同人使用nai训练，在anything系列模型上使用也很好，方便测试你的角色模型，选择你喜欢的吧，后续会更新各种风格,style05之前已经发过，style09很擅长nsfw,,style11需要使用chibi或者将分辨率设置为512x512
These are some common style models,trained by different people using nai, it is also very good to use on anything series models, it is convenient to test your character model, choose what you like, and various styles will be updated later,I have posted style05 before, style09 is good at nsfw,style11 needs to use chibi or set the resolution to 512x512
Description
FAQ
Comments (9)
Hello to the best and my most favorite civitai lora/model uploader in the whole fcking planet! Which extension do you use to create images with 2 distinct characters (like Tachibana Arisu and Momoka in the sample images)? I don't see "AND" syntax in the prompt that "Latent Couple" extension does and used by most people here in civitai (I think). Oh and I use "Latent Couple" extension but it doesn't work for me no matter what I do, I always get a single character in the image with the merged properties of 2 different character loras. If you don't mind me asking of course.
Oh and Happy New Year! Thank you for all the good and high-quality stuff you provide in the community.
Also, I wish you a happy New Year! In fact, this is simply due to the appearance of a large number of specific characters in the material. At the same time, the NAI model has a stronger effect on restoring characters. As for some of the dual-character roles I created, they also involve extensive training with images featuring these paired characters.
@kozue I recognize this artist it was based off from but I forgot his name ^^;
Yeah he often draw these characters from this series/game and it looks very adorable. Thanks for the explanation. But just in case you're wondering what I'm talking about the Latent Couple, here are some examples I uploaded: https://civitai.com/posts/1104257 (the characters may look "thicccccccccccc" which is kinda out of my control sadly)... lol
Once again, thank you for 2023! I hope to see more high-quality stuff you upload this year 2024 too.
@ddoscv Generally, when working with images of two individuals, I prefer to train them together as a single model (LORA) to achieve optimal results. This allows the AI to recognize each person separately, enhancing the natural interaction between the two. Of course, this is contingent upon having a sufficiently diverse dataset. Once again, I wish you a Happy New Year!
Brother do you have the SDXL version of this style21 (@yukkieeeeeen)?
may be train later
That'd be awesome, thank you.
My man! Another happy addition to my collection, thank you so much!