LoRA will be updated constantly to improve results, I highly reccomend using 3:4 ratio on portrait + Euler A sampling
Easy to trigger the position withouth breaking the rules / making sensitive content, safe to use for SFW and NSFW pics
Version 1.1 fixed random hand/arm movement through the different generated images, results are now more consistent with less nonsense
Version 2.0 has more pose variety compared to 1.0 and 1.1, now it has a high variety of poses to meet your requirements! I also expanded the database of LoRA while creating many outfits, characters and background scenery, so it is now possible to create as many variations of pics as you can. Every not mentioned pose in prompts now won't occur, it is now very strictly following your prompts. It is also working well with shoes without adding additional limbs etc. It is now also enhanced to better expose panties hidden under pantyhose, so now you can see more~ hehe :3
Recommended prompts to use in "negative" box: score_6, score_5, score_4, source_pony, (worst quality:1.2), (low quality:1.2), (normal quality:1.2), bad anatomy, unnatural body, error, extra limb, missing limbs, bad feet, extra toes, interlocked fingers, extra fingers
