Premium character LoRAs
Original synthetic characters, bundles and AI tools are in my shop:
Join the community on Discord: https://discord.gg/nB79wRNtMT
NSFW Outfits
A concept LoRA for Krea 2 focused on revealing outfits: sheer, cut-out and see-through dresses, lingerie sets, micro skirts, mesh tops, harnesses and heels, in mirror selfies and full-body shots. It gives an amateur photo look.
Usage
Start with the trigger word OutfitsRK, then describe the outfit precisely (garments, cut, fabric, sheer or cut-out parts, color, accessories, shoes) and what it reveals, then her skin, hair and makeup, her pose and framing and the place, and end with the light. The more precise the outfit, the stronger the effect.
OutfitsRK, Photograph of a woman with <skin> and <hair>, wearing <outfit: garments, fabric, sheer / cut-out parts, color>. She is <pose / framing> <place>. The lighting is <light>.
Recommended weight: 1. Made for Krea 2 Turbo (8 steps, CFG 1).
Prompt Examples
"OutfitsRK, Photograph of a woman with fair skin and blonde, wavy hair. She is wearing a tight, white, crop-top with cut-out sides that reveal her midriff and small, white thong with cut-out sides that expose her buttocks and vulva."
"OutfitsRK, Photograph of a woman with light olive skin and long, straight black hair. She wears large, black-rimmed glasses and a sheer, nude-colored bodysuit that is transparent, revealing her medium-sized breasts with small, pink nipples and her vulva."
"OutfitsRK, Photograph of a woman with fair skin and long blonde hair in a ponytail, wearing a cow-patterned, sheer, black-trimmed lingerie bodysuit with a front zipper and black garter straps attached to black thong panties. She has cow ears headband with pink bows on top of her head and a blue and white bow tie around her neck."
Training
- Base model: Krea 2
- Type: Concept LoRA
- Steps: 4000
- Rank: 32
- Optimizer: adamw, constant learning rate
- Engine: ai-toolkit
Generation settings
Trained on Krea 2, and works on Krea 2 Turbo for faster generation.
- Raw: ~30 steps, guidance 3.5
- Turbo: ~8 steps, guidance 1.0
The gallery images were generated with krea2_turbo_mxfp8.safetensors, at 8 steps, guidance 1, LoRA weight 1.