<h1><u>Sojourn loRA:</u></h1><p>This is my new loRA in the Overwatch universe, this loRA represents Sojourn a heroine of this universe!</p><p></p><h2><u>Trigger words:</u></h2><ul><li><p><code>Sojourn, Sojourn outfit, blue eyes, whole body</code></p></li></ul><p>The <strong>whole body</strong> is important if you don't want to do only close-up face because the loRA tends to do that quite regularly</p><p></p><h2><u>Steps, CFG scale, </u><strong><u>CLIP skip and loRA weight:</u></strong></h2><ul><li><p>The loRA is quite comfortable with a <strong>Steps</strong> between <code>20 and 45</code></p></li><li><p>I recommend a <strong>CFG scale</strong> between <code>5.5 and 7</code> for optimal results</p></li><li><p>For the <strong>CLIP skip</strong> <code>1 or 2</code> look pretty decent.</p></li><li><p>The weight of the loRA works best between <code>0.7, 0.8, 0.9 and 1</code> for best results. (<strong>note that your results may differ depending on the model you use, the VAE you choose, your prompts, the weight of your loRA..., you will probably not directly have the same results as mine)</strong></p></li></ul><h2></h2><h2><u>loRA training:</u></h2><ul><li><p>The loRA was trained with the template: <a target="_blank" rel="ugc" href="https://civarchive.com/models/46291/esuna-overwatch-girls">https://civarchive.com/models/46291/esuna-overwatch-girls</a>,</p></li><li><p>Model creator: <strong>Esuna</strong></p></li></ul>

Files with SHA256: 69c9f4ce835a05d97d7f41a720c896876483c9f7024e254802934ac747e3f917