Model trained on NoobAI-XL V-Pred-1.0 to imitate the style of Hsuliherng (へるんぐ / Herungu)=====Support the artist: <a target="_blank" rel="ugc" href="https://www.pixiv.net/en/users/2079339">https://www.pixiv.net/en/users/2079339</a>=====I've also tagged images with their OC using: <code>hsulichar</code>. It wasn't really a focus, so it is inconsistent, but it can get a close approximation of the character with some help + tries.To help with generating the character, I use the following:<code>brown hair, blue eyes, two side up, short twintails, white ribbon, hair ribbon, blue collar, belt collar, leash</code>NOTE: All the images I generate for the showcase are done with base NoobAI-XL without the use of any Adetailer or Inpainting, only Hires. fix to do basic upscaling.My results do not look as good as you can get quite easily with more stylized checkpoints or better finetuning, but I wanted to show the baseline (and that the LORA works). Try out what works best for you!

Hsuliherng Style (NoobAI vPred)

Hsuliherng_Style_V2(NoobAI).safetensors

v1.0

V2.0

Model trained on NoobAI-XL V-Pred-1.0 to imitate the style of Hsuliherng (へるんぐ / Herungu)=====I've also tagged images with their OC using: `hsulichar`It is not the greatest and needs a lot of help with tags, but it gets a close approximation of the character.To help with generating the character, I used the following:brown hair, blue eyes, two side up, short twintails, white ribbon, hair ribbon, blue collar, belt collar, leash

Support the artist: https://www.pixiv.net/en/users/2079339

Several changes and (I think) improvements! Updated the dataset and removed the need for the trigger word. I think this better captures the the "softness" of the artist's style over the previous version.

It is also slightly better at generating their OC (same trigger word as before: `hsulichar`), but it is still not 100% at it, so play around with it and expect it to take many tries to get it right.


Files with SHA256: abdd6bbc6f297729f56712b0523d92163a3288f5e11c4dff6e504af57f2dcdf2