After a freaking hell of study, testing and more testing UI's and hyperparameters here we are! This is my 1st project so don't be so hard with me :)

This lora is expected to work for all kind of content and with a good range of variation from original character (see examples). This is my 1st big success after all the learning and training i've done these days and i consider it very decent honestly.

This lora works on the bast majority of models with base in 1.5, but works really better in models with completely flat/anime style and coloring (like Project-K, CuteYukiMix, Counterfeit V.3, OxygenMix...) and also works pretty decent with these more volumetrics/2.5D (like Hassaku, CitrineDreamMix, SakushiMix, Kizuki...), but if you can choose or just don't care just stick it with the anime style.

DONT really need trigger words, but quality rise if you put "Nahida" in the prompt, and also if you want her original little size consider putting "child" or "small". (Or you can simply put Nahida with other contexts and get an "adult" versions of her if you want too).

If you like my work please consider Buy me a breakfast! :)
https://ko-fi.com/seikkv

Every donation will be destined to my AI learning, work and computer to upgrade all my work! Thanks!