Furry_Detailer (NoobAI XL)

”Her eyes are round and big!”

"This LoRA is a Generic Furry Maker."
For those who want to enjoy furry girl in a somewhat casual way.
No specific design, copyright, or character exists. It can be used for generic purposes to create your own original character.

Create your own furry girl!
Combining multiple furry type LoRAs is key.
Set all LoRAs to a low weight (0.3 or so) and increase the values of the ones you like.

Trained by NoobAI-XL (NAI-XL) model.
Use for various Illustrious models. (and official haruka/hoshino models)

This LoRA adds furry details to your image.

It can be used alone, but was created for use with other LoRAs.

It’s used 100 high resolution 2048*2048 images, and then all of them were manually retouched and adjusted.

The data set has been redesigned to create round eyes and chibi.

This LoRA is free for all to use.
Since this LoRA was added by me in my own handwriting, it is not in the style of any particular artist and can be used freely without worrying about copyrights.
Also, by using this LoRA, it would be possible to avoid too much resemblance to a specific artist style.

summary

human face with round eyes + white furry body edition.

It can enjoy changing human characters into furry, or combining them with furry LoRA.

Recommended LoRA strength value 5 to 8.
The lower the value, more human-like, and the higher the value, more furry.

About the trigger word
The concept is to add fur, so “body fur” is used. Adjust accordingly, if it needs more fully, rewrite this tag (body fur:1.5) etc.

Remove the "furry" tag if you don’t want to make it too furry.
(For example, when you want to make human face with furry body characters).

"round eyes"
You can also add a "round eyes" tag if you like. That would make them bigger eyes.

about ”chibi” tag?
"chibi" tag can be added to change the picture significantly. (very cute).
If you feel that her head is too big, put the "large head" tag in the negative.

How to low cost credit generate at this site
There are LoRAs that can significantly lower sampling steps. This allows for free or low-cost testing. [Hyper SDXL (link)](https://pixai.art/model/1783699585926255419) Using this LoRA together, set sampling steps to 8 and CFG to 2 for test.

Costs can be dramatically reduced. (but would instead degrade image quality). Please try to generate free images with this first, as they may not be exactly what you expect. Don't waste your credits. There are many other LoRAs that can reduce generation costs, so please find the one you prefer.

Use as reference my artworks
You will find examples at the bottom of this page under "Artwork made by this model".
Most of my examples are based on used "low cost LoRA". Please refer to them.

note
This LoRA can be used to detail changer, body shape, eyes, etc.
It’s a good idea to combination using other detail types and background LoRA.
It’s also good idea to use this LoRA with other furry type LoRAs.

Recommended to set some clothes.
Models often perceive fur as clothing, and the tag in this case should be emphasized. (shorts,shirt:1.3) etc.

Recommended to set some background tags. If not specified, the background will be white.

Author's Comments.
Again, the base dataset was so fluffed.
However, it did not turn out as expected. :<
Instead, the eye detailing seems to have worked as intended. :p

I use it regularly myself, but don't worry about it for use this LoRA. Even if the exact same image is uploaded, it's more fun that way. Examples of my artwork are for copying settings. Feel free to use them and upload them.