Home Artists Posts Import Register

Downloads

Content

I'm hungry, I want a snack 🍼

🖼️ See some examples 🖼️

⚖️LoRA prompt weight: 0.8~1
⚙️~40% effectiveness
🏫Model used for training: personal modded model

I know, the previous model was V05. I put the version of the previous model like this to show how many versions I had made previously. If I followed that line of thought, that model would be V34. 

In addition, more than 20gb of intermediate LoRAs were produced to achieve this result.

🔍Recomendations:

  • CFG ~9 ⭐
  • Steps >=35
  • DPM Family

I recommend start your prompts with:

jkbfeed, 2girls, one_woman_lowered_making_jkbfeed_on_another_woman, ...

LoRA details and tendencies (+resolution):

  • I achieved good results with these neg embeddings: CyberRealistic_Negative-neg and FastNegative. But you can use normal neg prompts.
  • You can use viewpoints prompts, but this don't have a great impact on results.
  • This model have a great impact on generations, as result, faces and hands sometimes break. Use hires or adetailer to correct errors.

Inside the folder I left some example images, you can use this on your prompts.

Comments

JKCHSTR

The big problem with making this model is that UNet, SD's image-generating system, thinks that a woman sucking a breast is someone sucking a penis or kissing someone else. I had to trick the system into accepting this concept.