Original Author text:
Proudly exhibiting his newly captured meat shield, just recently strapped on his armor, our monstrous humanoid decides he might as well have some fun along the way...
Trained on Anything v3 - so I guess it will work best with models using booru - style tags.
Activation words: onarmor, meat shield, fanny packing. I recommend Clip Skip 2.
"onarmor" was chosen in an attempt, not to weaken "real" tags, but I guess I don't fully understand, how Lora-learning works, so this might have been the wrong approach.
Depending on the model, you might want to give more weigth to the activation words (onarmor:1.2), (fanny packing:1.2)
The lora mostly delivers the pose. To acutally see some action, add some "sex" activities, bodily fluids, and see what SD makes of it...
If you happen to use "meat shield" in the prompt and literally stumble across pieces of "meat" in the output, I guess I this is either because I didn't really "get" how activation words work, or that my Lora is incompatible with the used checkpoint/model.
My first Lora. Grabbed a whole bunch of images from the 'boorus and gave it a go. Next time, I guess, I'll try less pictures, but better quality/curation?