Premium character LoRAs
Original synthetic characters, bundles and AI tools are in my shop:
Join the community on Discord: https://discord.gg/nB79wRNtMT
Cum Covered
A concept LoRA for Krea 2 focused on cum shots and facials: women with white, viscous cum on their face, mouth, tongue, breasts and body, from a light touch to heavy coverage. It gives an amateur, candid photo look, in bedrooms, bathrooms and living rooms, topless or nude, often smiling or with the mouth open at the camera.
Usage
Start with the trigger word CumSlutsRK, then describe the woman (skin and hair), her body and breasts, what she wears or that she is topless or nude, where the cum is and how much, her expression, her pose and the place.
CumSlutsRK, Photograph of a woman with <skin> and <hair>. She is <topless / nude / wearing ...>. She has <a small amount / a lot> of white, viscous fluid on her <face, mouth, tongue, breasts ...>. Her expression is <expression>. She is <pose> <place>.
Recommended weight: 1. Made for Krea 2 Turbo (8 steps, CFG 1).
Prompt Examples
"CumSlutsRK, Photograph of a woman with medium brown skin and long, straight black hair. She is topless, exposing large, natural breasts. She has cum on her face, mouth, tongue, and breasts. Her expression is a smile, and she is looking at the camera. She is kneeling on a wooden floor."
"CumSlutsRK, Photograph of a woman with light skin and brown hair, tied back in a ponytail. She is topless and wears a brown skirt with white polka dots. She has a small amount of white, viscous fluid on her face, chest, and breasts."
"CumSlutsRK, Photograph of a woman with fair skin and bright orange, shoulder-length hair. She is wearing a pink, frilled, strapless bra pulled down and black thigh-high stockings. She has a small amount of white, viscous fluid on her face, tongue, and chest."
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.