Remember the golden age of eroge? This LoRA brings back the look of mid-2010s Japanese visual novel CGs — crisp clean lineart, soft glossy shading, and those unmistakable 2015-era anime eyes.<h2 id="trigger-word">Trigger word</h2><pre><code>nijigen2015</code></pre>Put it at the very beginning of your prompt.<h2 id="how-to-use">How to use</h2><ul><li>Base model: Animagine XL3.1 — other SDXL anime models may work, but this is what it was trained on</li></ul><ul><li>Weight: <code>0.8</code> – <code>1.0</code></li><li>⚠️Do NOT use year/quality tags like <code>newest</code> — they push the style back toward modern anime and fight the LoRA</li></ul><h2 id="training-details">Training details</h2><ul><li>SDXL LoRA — dim <code>8</code> / alpha <code>1</code>, 20 epochs</li><li>109 images, captioned with WD14 tagger</li><li>Full training parameters are embedded in the file metadata — check the details panel on this page</li></ul>

Files with SHA256: 310c6ada942a1ba7efaa17498825e7079bf9da04f2c6c4c1e3a4a2778f0d39a2