上杉达也は 世界中の谁よりにも 浅仓南を爱していま
上杉达也爱浅仓南，比世界上任何人都爱
这个模型使用了超过250张高质量的原始图片进行训练。
这个模型拥有很好的画风模仿效果。训练图片全部来自该画师的原画集。
Flux版本：https://civarchive.com/models/810366
illustriousXL 版本：https://civarchive.com/models/1399783
Pony XL版本：https://civarchive.com/models/374837
推荐权重：0.6~1.0
推荐CFG：5~6
推荐模型：Animagine XL V3
--- --- --- --- --- --- --- ---
This model used more than 250 high-quality original pictures for training.
This model has a good effect of restoring the artist's style. The training pictures are all from the original painting collection of the artist.
Flux version：https://civarchive.com/models/810366
illustriousXL version：https://civarchive.com/models/1399783
Pony XL version：https://civarchive.com/models/374837
Recommendation Weight：0.6~1.0
Recommendation CFG：5~6
Recommendation Model：Animagine XL V3
Description
FAQ
Comments (4)
Many thanks! Adachi is without a doubt one of my favorites.
Out of curiosity: which base model are you training your style Loras on? I get great results with the one you recommend (Scmix) but the moment I move to other anime models things doesn't look so good.
My lora models are all trained based on SCMIX_NSM. Because SCMIX_NSM itself is a mix model, not all animation models can have better results.
Please try : pastel, abyssorangemix3AOM3, meinamix_meinaV8, Night Sky YOZORA Style Model
Any chance you would consider releasing at some point an alternative version of your LORAs retrained on the original NAI model (animefull-final-pruned-fp16)? Because that is the "grandparent" for all the other anime models, I think that would really increase the compatibility of your style LORAs.
Many thanks in any case and keep your great work!
@MelmothTheWanderer
Let me answer this question.
Although NAI is the most compatible, it is not the most effective.
I used NAI as the base model, used the same training materials, Tag information, and the same training parameters to get a new model. I randomly generated 2 pictures and updated them in POST. I believe you will understand the reason why I did not choose to use NAI as the base model after reading this.
Thank you very much for your support~