My third model. Honestly, I do not know for who it can be useful, because it is literally a hole from which the tongue sticks out. But maybe it's just my lack of imagination :)
v1.0
Using some good prompts, u can get really incredible results. For example, in dataset of model zero images with eyes of girl, only one tongue, but if u simply prompt "eyes", it's significant increase chance, that picture will contain whole face of girl. Also this model works well with my another model "Open Gag Mouth"
Tips: 0.6 - 0.8 scale will be good (with some LoRas need higher or lower)
Training model (also used to generation): abyssorangemix2_Hard
Keywords: glory hole tongue, pov, cum on tongue, salva on tongue, open gag mouth (last tag maybe bad working, so pls for this option download my another model Open Gag Mouth)
v2.0
Using this LoRA, you get an image with the face of a girl who sticks out her tongue. By adding the prompt "cum on tongue" you add cum to the tongue and some parts of the glory hole (not always). Works with character LoRA's (not good with all, for some LoRa's u need to lower scale), the example shows LoRA Lucy from Edgerunners and Prinz Eugene from Azur Lane. Maybe it's a little overbaked, but what do you want from a hole with a tongue in it? It was also quite difficult to assemble the dataset, even the one on which this LoRA model trained.
Tips: 1 scale will be good (with some Character LoRas need higher or lower)
Training model: NAI
Model used to generation: abyssorangemix2_Hard
Keywords: glory hole tongue (also u can add "cum on tongue")
If u know how to improve model, let me know.
Description
Arrived
FAQ
Comments (19)
does it work for realistic models?
Hi, i'm not tested this, but dataset all contains anime style. U can try it and give feedback if u want
Error i cant start it :I
@DarkSide675 what kind of error u get?
it might have been a bug with the website, try re-downloading it, might have had the same problem
i did it worked @worgensnack Thanks
@DWeoss But now i can only create black images tried different models but only black images again
@DarkSide675 I downloaded the model directly from the site and tried to generate it - it works for me. To be honest, I don’t know what is the reason for this behavior of your AI, maybe you have some settings set, or it turns out that you are downloading a broken file. What sampling do you use, how many steps, what scale cfg, image size, seed, model scale?
@DWeoss No more error just a black image i tested it with multiple models all are just black
im new here i can post images?
@DWeoss Let me test it with other settings
@DWeoss Now it works . Deleted all other stuff
@DarkSide675 Can you please specify what exactly did you remove?
@DWeoss idk what was causing the Error i only know i deleted other Loras some models and the embeddings
@DarkSide675 Anyway, glad to hear that everything was resolved by itself.
@DWeoss I guess i should reinstall my AI after that error :D
EDIT: I redownloaded version 1.0 and it works!
Sadly, this Lora crashes my automatic1111. gives me a crazy NaN tensor error and I have to restart the whole thing. bummer, this looked very cool.
@Orbo Dude, don't worry, I'm going to make a new version of this LoRA soon, using a different learning model and different settings. So you can just wait and maybe you can enjoy it later :)
FYI, I had the same problem, then I found I didn't use the latest version of automatic1111, I updated, and the problem was solved. Maybe this helps.
@DWeoss Redownloaded it and it works! great job!