I was curious what would happen if I trained a lora on a huge amount of data, for this lora I collected 2900 images from all seasons of anime.In most cases she will be in her clothes, but for stability you can use: <pre><code> black dress, thigh strap </code></pre>In a negative promt I recommend using:<pre><code>star \(symbol\), </code></pre>

Files with SHA256: 1b2868b8e1449a871e3db5d1776be5f2b1ef98a23c2b5a0ab99dcaa7c3c77288