사용자 LoRA

XL

Futa on female POV

#성숙한 여성

#1인칭 시점

#소녀

#후타x여

커버 제거됨: 콘텐츠 규칙 위반

-커버 이미지를 다시 업로드하십시오

  • 여전히 이 모델을 검색하고 사용할 수 있습니다
업로드 날짜
2024. 5. 25. PM 4:36
이용수
815
리뷰
훌륭함  (1)
트리거 단어

fpv

권한
  • 이미지 생성 및 공유 허용
  • 다른 사용자가 내 모델을 다운로드하도록 허용합니다
  • 생성된 이미지의 상업적 이용 허용

설명

Futa POV: a shot in which the viewer is a futa doing various actions.There are more of these floating around, but the usage of those has been finicky at best. Here's my attempt at making one that is slightly more consistent and yet has variety.V2 is trained on 72 images using anylora for 5 steps at 55 epochs.Instructions: Set at strength 0.4-0.7 for v2, 0.75-1 for XLv1 (probably works at other strengths but haven't tested it). Feel free to use any base model you feel like, but it is recommended to use futanarifactor. Note: the futa is counted as a girl so having 1 other girl in view means using 2girls in the prompt. Using 1girl and solo together in the prompt should generate only a futa pov.Use the following trigger tag: fpvAmount of girls: 1girl+solo, 2girls, 3girls, 4girls, 5girlsOptional tags: veiny penis, foreskin, fellatio, sex, vaginal, paizuri, cooperative fellatio, motion lines, saliva, precum, heart-shaped symbols, heart-shaped pupils, pubic hair, holding head, Experiment!Gacha: lipstick marks, kiss marks, penis on cheek, panties aside, close-up, imminent sexXL specific: highly recommend the following in your prompt: (fpv), futanari pov, futa pov, (own breasts). V2 is proudly less gacha than V1. Enjoy a bigger variety of poses and consistency.XL has better capacity to generate scenes with multiple girls at the cost of some inconsistency. You will not always see your own breasts. Try to stick to the recommended prompt above if you have trouble getting what you want. If you see any artifacts (happens) then lower the weight of the lora.Later versions may be more consistent and having less trouble with artifacts and more options for dressing your own futa seperate from the girls displayed (surprisingly difficult to get to work). Feel free to send good images for a dataset my way.

댓글