Send me buzz, I'll make more of these, and better quality, too.
NSFW posing and touching LoRA for MiniMax H3. Triggerless. No activation token.
I2V PROMPT:
"For the target video, at 0.00 seconds into the target video, <Picture 1> (from [Shot 1]) is fully referenced.
integrated_multimodal_description: [Shot 1] Ultra photorealistic high definition sexy webcam video of a very beautiful asian filipina young woman from the first opening frame, maintained exact in appearance throughout the shot without drift. Her face and body is highly detailed, featuring a curvy-fit sexy body that matches her body frame with authentic and natural body proportions, She has photorealistic natural skin texture, and perky, very firm, round augmented breasts that feature small detailed nipples. The photorealistic lighting aligns with first frame light source directions, maintaining the same lighting direction, with correct and photorealistic light source-derived shadows. Natural subject motion with believable articulation and muscle tension, physically accurate textures and materials, and life-like facial micro-expressions. The subject is kept fully recognizable at all times."
TIP: Try BJ / HJ scenes. Worth it.
Trained on softcore and midcore videos of attractive women posing and touching, with emphasis on round, full breasts, hand contact, and torso presentation. Works on T2V, I2V, and FL2V. Recommended checkpoint: epoch 10. Strength 1. 10 epochs, 10,000 steps.
What it holds
- Round, full breast shape and weight through pose changes and hand contact
- Hands on hips, hands lifting or cupping the breasts, fingers along the outer curve, palms on the upper thighs
- Side-to-side torso and hip shifts that keep the chest in frame
- Identity, tattoos, clothing, and scene from the reference on I2V and FL2V
Use
- Leave the prompt triggerless. Do not lora trigger; prompt naturally.
- Name the touch. "Hands rise and slowly cup the breasts" beats "poses sexy."
- On I2V, lock the first frame, then add only the motion. On FL2V, point Picture 1 at 0.00s and Picture 2 at the end time.
- One continuous shot. (if you choose.)
- Keep strength at 1. Drop toward 0.7 if hands melt or breast shape drifts off the reference.
Example, I2V:
For the target video, at 0.00 seconds into the target video, <Picture 1> (from [Shot 1]) is fully referenced.
integrated_multimodal_description: [Shot 1] Absolute frame-to-frame continuity holds on the woman, her identity, facial structure, anatomy, proportions, and established appearance unchanged. The camera stays locked off, with perspective, scale, framing, and lighting unchanged. She stands nude in a relaxed stance, dark hair over her shoulders, round full breasts, long light nails, hands resting at her outer thighs. From this exact first frame her hands rise and slowly cup the outer curves of her breasts, then she turns slightly side to side, gentle torso and hip shifts keeping the breasts revealed and round; hair shifts with the body, and skin and nails keep natural weight and contact.
overall_soundscape: Quiet room tone, soft breathing, and the light sound of palms moving over skin.
non_diegetic_music: N/A
Example, FL2V, hands from hips to thighs:
How the reference pictures align with the target video — Picture 1 (from Shot 1) aligns with the 0.00-second mark of the target video; Picture 2 (from Shot 1) aligns with the 6.00-second mark of the target video.
integrated_multimodal_description: [Shot 1] She begins as in Picture 1, hands at her hips, round breasts presented to the locked camera. Through one continuous take her hands slide down the waist and settle on the upper thighs as in Picture 2, with a slight side-to-side torso shift between the two poses. Breast shape stays round and full. Identity, tattoos, and the room do not change.
overall_soundscape: Quiet indoor ambience, soft breathing, faint skin contact.
non_diegetic_music: N/A