Z-TURBO_BRK_SPANKING

BRKN-Spanking & Bent-Over Poses LoRA

LoRA focused on natural, over-the-knee spanking and bent-over poses between consenting adults. Trained on ~40 images over ~5,000 steps, it captures classic disciplinary setups (over the knee, bent over a chair, over a desk) with a slightly sensual, stylized tone while keeping the environment flexible.

---

What this LoRA does

• Core poses

  • Over-the-knee spanking, natural body weight and balance

  • Bent over a chair

  • Bent over a desk or table

  • Standing / leaning variations flowing out of those main poses

• Camera angles

  • Front view

  • Back view

  • Three-quarter view

  • Works both for wider scenes and closer framing on the pose / impact area
    Works also on super close-up rear

• Character setups

  • Works with male/female and female/female pairings

  • Designed for clearly adult characters only

  • Can be mixed with character LoRAs as long as you keep the weight moderate

• Visual details

  • Visible hand / slap marks after impact (redness, impact prints), Hand raised awaiting

  • Natural body curvature when bent over

  • Can also handle close framing on the bent-over pose / lower body position

---

Training info (for the nerds)

• ~40 training images

• ~5,000 steps

• General OTK + bent-over discipline theme

• Includes variations over chair, desk, and classic lap positions

---

Recommended settings

• LoRA weight: ~0.9 – 1.1 (start at 1.0 and adjust)

• Sampler: Euler / Euler a (simple is a safe bete but res_2s is also great for extra detail settings)

• Checkpoint: works best on your “beta / res2, second set” style models or similar semi-realistic checkpoints

---

Example base prompt (safe starter)

BRKN_SPANK, SPANKING , SPANK

(BRKN spanking, bent over knee), over a chair, over a desk, natural pose, two consenting adults, slight redness on skin, dynamic three-quarter view, soft studio lighting, detailed anatomy, cinematic composition.  

ex: Lay across his knees , lay over his knees

You can then add your usual style stack (model name, lighting, camera, outfit, etc.) on top of that and just plug your LoRA tag at ~1.0.