Sign In

【REVIVAL】Lico Project | Lora Style

98
886
4.3k
29
Updated: Jul 17, 2024
styleanimesexywomanfurrygirls
Verified:
SafeTensor
Type
LoRA
Stats
211
1,846
Reviews
Published
Jul 1, 2024
Base Model
Pony
Training
Steps: 10,433
Epochs: 10
Usage Tips
Clip Skip: 2
Hash
AutoV2
8A50E3A0DD

Hello welcome!

Here I leave a small introduction of the LoRA:

Revival - Ultimatum | V1.0:

So... the main idea was to delete the images that were added by mistake and in the end I ended up doing Metanoia, yes, I still have the data but is it necessary?

I decided to improve the faces and mix the style much better, that breaks the 2d (unless you add it to the prompt) but it looks good and I also added a chromatic aberration effect, I suppressed the Lico_PS trigger for a better fusion but I recommend using it if you want it to be the base of the style (it is not erased, just scattered).

This style only fixes the Ultimatum problems, it still has the same data:

cyclops, furries (birds, anthro, etc), humans, monsters, etc.

All of that is included here, of course, there are also ponies, but the most important thing is that it looks good, yep~ it's not a big deal but it's important to me, I'm very picky about the results and well, greetings ~ XD


REVIVAL IN PROGRESS! - Ultimatum | V2.0:

Lico has become a thorn in my side...

I started this thing as "Lico" and it came out average, then I continued with "Lico Project" and it came out something very far from what I expected, almost 70% of the data was discarded because of a couple of styles that were not even part of the 5% and that frustrates me.

I mean, if it had been my intention for that merger to occur I would accept it but it was a pretty stupid mistake that I only realized when I went to review the images to see the prompts for the data set.

This will be the last version for Lico, this will be his rebirth, this will be his REVIVAL!


BETA - Ultimatum | V2.0:

If we're getting picky, let's say Lico was a 2.5d style, but it's not!

I made a MISTAKE and added some images I made with T-Ponynai that altered my entire 2D style and well... I trained 6 versions because I didn't think I made that mistake, I'll just leave the third one in a zip and tomorrow I'll upload it to sandbox, by the way I also uploaded a new version for Apocryphal ~

And I promise to improve this style XD


This style works on t ponynai | V1.5:

I'm working on Lico, I made a mini furry version but I'm editing images (removing signatures and censorship).

This time it will be a full style and I will add a main trigger but I don't know how long the training will take, if all goes well it will be out in a few hours, if not it will be released tomorrow.


This style works on t ponynai | V1.0:

I was planning to train a lyCORIS, in the end I completed my dataset and put it to train like Lora, yes... I forgot.

It has furry art and is very different from what I expected to get, there are also many examples of bird girls (I don't know why there is so much river porn, I know that there was a time when people got angry with Jewel for Blu but... forget it, It doesn't matter, I'm just too lazy to upload it to tensor to get the examples, I'll wait until it's loaded and I'll generate them here...)


PD: It doesn't work, I mean, if it generates images but it didn't mix well, I'll fix it for later...although I knew it would happen since it uses little data, I honestly just wanted to be vague (that's why the size is so small).

PDv6