<h1 id="overview-nao1hz0pw">Overview</h1><p>This model is intended to produce side by side pictures of the same woman before and after a cumshot. It is also quite capable of producing cumshot / facial images without the before.</p><p></p><h1 id="prompting-rb17stl4g">Prompting</h1><p>The LoRA expects a very specific type of prompt - take a look at the showcase prompts.</p><p></p><p>Examples:</p><pre><code>side-by-side before and after cumshot photos of a 20yo red haired woman with freckles, on the left side she is wearing a red plaid shirt. On the right side is a close up with her face covered in semen which is a thick, viscous white substance. The appearance of semen suggest she received a facial cumshot. On the right side she is naked and pushing her breasts together.</code></pre><p></p><h1 id="training-3bieglqs6">Training</h1><p>This model was trained using the excellent <a rel="ugc" href="https://github.com/ostris/ai-toolkit">ostris/ai-toolkit</a> using the default training configuration, just increasing the step count.</p><p>I used 18 images, I was planning to use 120 that I collected but I did a test run with 18 to check parameters and it turned out so well that I published it.</p><p>The captions were in the style of the example above which I did manually.</p>

Files with SHA256: e5a166e2a9162427a6a36c9d6ebd23ae069766e394cb6357bf2f0cb7459d0932