Better trained lora that has better positions and viewsedit for minimax H3: there are no triggers for this. Been testing on site and found that T2V and I2V work out real nicelyFollowing Trigger words are a bit spotty, MGOnaholeposition, MGTopviewOnahole, MGLowviewOnahole, MGReverseOnaholeposition, MGSideviewOnaholeedit for V4: for the #handheld# triggers number is how many hand holding the minigirl, end letters for where the hand is grabbing, T for torso, L for leg, A for arm, the S and T in 1#hanhelda is for one hand holding one arm or together in one hand. it a bit of hit or miss but now might have more controle on where the giant hands areV5- add a few more images and tweaked a few stetting on bakingprompts like side view, low angle, and top view can help with generation. the suggest resorce can help with some triggers but I have a few failures so test to see what works. see the top images for prompt help.Illustrious Ver- just to add so cause there was not alot of size differnce in Illustrious. Does not work well with characters loras. if you cant get the miniperson the righrt size use the format (miniwomen:#.#) and add weight. I found 1.3 to 1.5 works best

Minigirl Onahole (pony & illustrious)

first try on a H3 trained on site but it came out good.

minigirl_onahole_H3_minimaxh3_3383568_epoch_10.safetensors

Files with SHA256: 8ca9c88b6562704a9c483c451ad0d820598bafaa4440388cfd72a79b82ba5bf5