The theme of this LoRA is to treat small insects and other winged fairies as if they were masturbators!
In my environment, I was able to use it in combination with other LoRA characters to make them into fairies!
However, I am not sure if this LoRA can be used with all LoRAs.
Thanks to an informant, I have found out.
This LoRA is not suitable for a realistic system.
After hearing this, we decided to create a realistic version of this LoRA.
If you want to use this LoRA for realistic applications, please wait for a while!
It will take some time, but we hope you understand.
Description
FAQ
Comments (6)
We are very sorry, but we have had to replace quite a few images.
This is because, from the perspective of the patrolling AIs, all of the images look quite suspicious, so we have reluctantly decided to replace them.
I didn't mean to and didn't intend to, but now that I can see them, I have no choice.
My apologies to those who appreciated the original images.
It does not work with the majority of both realistic and semi-realistic models. It only works with a couple of hentai-focused models. Some prompts contain parts of XL. Is the lora for XL or SD?
Thank you for providing the operational information!
I think the reason for the lack of realistic model behavior is that most of the training data is two-dimensional.
There is not much training data for realistic models and we are in the process of creating and collecting them.
We intend to release a realistic version in the future.
And about the prompts, these are for SD.
I have included some prompts for XL because that is where I have had the best results in terms of learning in my environment.
@darknightnk716410 Ok, however the xl parts in the prompt are not functional for SD. I also tried 2D models but there is no way to make Lora work. I do not use the generation of Civitai, but only the one in local on PC. I tested Perfect World and Hassaku but I don't have your results even by putting the same prompt. In all cases I have images consistent with the prompt but also by eliminating the Lora in the prompt I have the same result, so it doesn't work in my case. I also tested the model used by you with poor results. Probably the model chosen for the training has influenced a lot making the Lora not flexible. This is because usually all the 2Ds also work on me in semi -seven models that I created since they are mixed mixes on purpose. It would be nice to have a more flexible version and I am waiting for future developments of the Lora. Thanks for your work.
@MyIA I am still learning about creating LoRA.
Any information you can give me would be appreciated!
As for the prompt, I may have misunderstood that it was working.
I am still trying to figure it out.
Thank you very much for the useful information!
@darknightnk716410 Where do you train? On civitai or do you use a program on the PC (eg: Kohya_ss)? Which model do you use for training?!