<h1 id="premium-character-loras">Premium character LoRAs</h1>Original synthetic characters, bundles and AI tools are in my shop:<a target="_blank" rel="ugc" href="https://rkrra.com/civitai">https://rkrra.com/civitai</a>Join the community on Discord: <a target="_blank" rel="ugc" href="https://discord.gg/nB79wRNtMT">https://discord.gg/nB79wRNtMT</a><h2>Cumshot Selfies</h2>A concept LoRA for Krea 2 focused on selfies with cum on the face or body: arm's-length and mirror phone selfies, playful expressions and tongue out. It gives an amateur photo look.<h3>Usage</h3>Start with the trigger word <code>CumSelfieRK</code>, then describe where the cum is and how much, her expression, how she holds the phone, her skin, hair and makeup, what she wears or that she is nude and the place, and end with the light. The more precise the cum and the selfie, the stronger the effect.<code>CumSelfieRK, Photograph of a woman with &lt;hair&gt;, taking a &lt;arm's-length / mirror&gt; selfie with cum on &lt;where&gt;, &lt;expression&gt;, wearing &lt;... / nude&gt;, &lt;place&gt;. The lighting is &lt;light&gt;.</code>Recommended weight: 1. Made for Krea 2 Turbo (8 steps, CFG 1).<h3>Prompt Examples</h3><blockquote>"CumSelfieRK, Photograph of a woman with fair skin and red, shoulder-length hair, wearing large, clear glasses. She has cum on her face, with some on her chin and a trail down her neck. She holds the phone at arm's length for a close-up selfie."</blockquote><blockquote>"CumSelfieRK, Photograph of a nude woman with fair skin and blonde hair, wet and slightly messy. She has cum on her face, chest, and upper body, with visible streaks. She is standing in a bathroom, taking an arm's-length selfie from above."</blockquote><blockquote>"CumSelfieRK, Photograph of a topless woman with fair skin and straight, light brown hair. Her face has visible cum on the right side of her forehead, cheek, and chin, with some on her tongue. She holds the phone at arm's length for a selfie."</blockquote><h3>Training</h3><ul><li>Base model: Krea 2</li><li>Type: Concept LoRA</li><li>Steps: 4000</li><li>Rank: 32</li><li>Optimizer: adamw, constant learning rate</li><li>Engine: ai-toolkit</li></ul><h3>Generation settings</h3>Trained on Krea 2, and works on Krea 2 Turbo for faster generation.<ul><li>Raw: ~30 steps, guidance 3.5</li><li>Turbo: ~8 steps, guidance 1.0</li></ul>The gallery images were generated with <code>krea2_turbo_mxfp8.safetensors</code>, at 8 steps, guidance 1, LoRA weight 1.

Files with SHA256: 1f5098614d9fc464aed2d4b6aa421dd891d4c6c017298c969be2cae2a948426e