Premium character LoRAs
Original synthetic characters, bundles and AI tools are in my shop:
Join the community on Discord: https://discord.gg/nB79wRNtMT
Face and Tits
A concept LoRA for Krea 2 focused on topless selfies framed to show both the face and the breasts: smiling or playful expressions, phone held up or mirror shots. It gives an amateur photo look.
Usage
Start with the trigger word FaceTitsRK, then describe her face (expression, eyes, makeup, hair) and her breasts (size, shape, nipples), how the selfie is framed (phone held up, mirror, from above), the place, and end with the light. Keeping both the face and the breasts in frame brings out the effect the most.
FaceTitsRK, Selfie of a woman with <hair>, <expression>, topless, her <breasts> in frame, <phone held up / mirror> in a <room>. The lighting is <light>.
Recommended weight: 1. Made for Krea 2 Turbo (8 steps, CFG 1).
Prompt Examples
"FaceTitsRK, Photograph of an adult woman with fair skin and long brown hair, slightly wavy, partially covering her right shoulder. She is topless, exposing her large, natural breasts with pink nipples and areolas. Her expression is a gentle smile, with light makeup, including pink lipstick and subtle eyeshadow."
"FaceTitsRK, Photograph of a nude woman with fair skin and long brown hair, sitting on her knees on a brown suede couch. She has a slim build with small to medium-sized breasts, pink nipples, and a visible tattoo of flowers on her right shoulder. Her expression is smiling with blue eyes, and she has minimal makeup."
"FaceTitsRK, Photograph of an adult woman with light skin and brown hair styled in pigtails. She is topless, exposing her medium-sized, round breasts with pink nipples and small areolas. She is seated on a white surface with her right hand resting on it, taking a handheld selfie from above."
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.



















