Hi there, here it comes another character lora.

This lora is aimed to help in the generation of images of Nino Nakano from quintessential quintuplets.

This time i can finally say that (after a few tries) this lora is on another level compared to the last ones i've trained.

This will be a shorter description compared to other ones.

---

Base version

It is a pretty much plug & play lora, that doesn't need any exceptional tweaking.

It was trained using mixed images from doujin, manga & illustrations, and i have to say that i am pretty pleased by this result.

It has a single activation keyword:

nakanonino

The hair ornaments are part of the lora and can be controlled by:

[color] ribbon  <-- example: "black ribbon" produces hair ornaments similar to the classic ones she wears

This lora support the generation of nsfw images, and can also be used to produce images much different from the context of the serie (i.e. the images of the armored nino, or the adventurer one in the gallery).

Even toh the dataset contained many images of heavy styled images the lora isn't that biased, and at weights around 0.75 can maintain the general style of the model used.

The suggested weight of the lora is between: 0.6 and 1.

The optimal weight is, in my opinion, at around 0.7 - 0.8.

There are few exceptions, more precisely, when pairying this lora with other loras like the expressions lora. In that case you might wanna increase this lora weight, or reduce the expression lora one.

p.s. The images in the gallery where produced using different models, all using euler a sampler, at 512*768 (exept one 768*512) resolution and then upscaled. (they were produced using adetailer, face + fulleyedetection).

If you like this model, you can support me by sharing the images you've generated here, making a review, or by tipping buzzs.