Hello welcome!
Here I leave a small introduction of the LoRA:
This week it was time to clean the PC and since I decided to delete hundreds of images, well... I'll give them a use before killing them ~ xD
But since there are so many images I didn't want many features to be lost and I will separate them into groups (something similar to my Dragon Ball compilation), I think 14 styles could come out of this and the first two came out well, Well. See you tomorrow!
PS: US is not EEUU, it is because of Únicos - Sueltos (I didn't really think about it much when I published it... although it is ironic because most of the artists are Japanese xD)
V1.0:
Unue | Ĉu vi konas Esperanton? v1.0:
For the first mix I added a lot of semi-realistic images, that means that if you add "eyelashes" the images will take a somewhat 2.5D route.
V2.0:
Due | Esperanto estas universala lingvo, kiu neniam funkciis v1.0:
If you used Google Translate to read the titles you're probably wondering why, well that's because I didn't think of good titles and here we are... Teaching?
For this version I increased the images from the anime and put shots with simple backgrounds.
V3.0:
Trie | Ĝi estis inventita de L. Zamenhof surbaze de la romantika, ĝermana kaj slava lingvoj. v1.0:
The use of data for this version includes an increase in quality (I ordered them from low to high), improvements in some perspectives and a couple of hatching styles XD
Description
FAQ
Comments (8)
sorry im really dumb and english is not my first languaje... do this mean that v1.0 and v2.0 are diferent loras?
I'm also bad at English. Yes, they are both different loras. XD
@Herrscher_AGGA2023 Thanks for the reply. wouldnt be better to call them something different than v1,v2,v3 ? like A,B,C or Iron, cuper, carbon ?
@Maxetto Yes, I hadn't thought about it, I'll change it, thanks XD
@Herrscher_AGGA2023 to you for making these loras!!
So this is a lora consisting of random images you generated. So there will be your images generated with random artstyle you used instead of our prompts?
It's trained just like any other style lora, it's just trained on images generated with style mixes to create a new style based on those.
Exactly, most styles are created like this, for example, if you like a type of eyes, but you don't like their bodies, you just cut out those parts and add other references, this is the basics of mixes.
That's what I did with Concupiscent, I really like Rororogi Mogera's type of faces, but I didn't find many images of full bodies and I replaced them with other authors of the same type. :3