<p>A portrait focused PonyXL style LoRA inspired by <em>Dungeon Meshi</em> race swap charts, following Kui Ryoko's art style.</p><p><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/1415a0e3-a0b0-4903-83b1-0d193fa84939/width=525/1415a0e3-a0b0-4903-83b1-0d193fa84939.jpeg" />Again, I post this LoRA mainly for fan art creation purpose. <strong>LoRA users should mangage their own copyright issues on generated pictures.</strong></p><p>Trigger base (optional): <strong><span style="color:rgb(250, 82, 82)">portrait</span></strong></p><h1 id="heading-812">Tall-men | 人間 | 长身人</h1><p>Although &lt;man&gt; in the trigger will influence the results quite a lot, I have used LECO (already merged into the published LoRA) to remove &lt;1boy&gt; from &lt;man&gt; so that the trigger can handle 1girl properly. Using &lt;human&gt; as trigger instead would be much straightforward but I really want to keep the original phrasing.</p><p>Trigger: <strong><span style="color:rgb(250, 82, 82)">tall-man</span></strong></p><h1 id="heading-813">Elves | エルフ | 精灵</h1><p>Trigger: <strong><span style="color:rgb(250, 82, 82)">elf</span></strong></p><h1 id="heading-814">Half-foots | ハーフ フット | 半身人</h1><p>I intentionally kept the space in the upper part of the portrait in dataset, but unfortunately that was not learned in the trigger.</p><p>Trigger: <strong><span style="color:rgb(250, 82, 82)">half-foot</span></strong></p><h1 id="heading-815">Dwarves | ドワーフ | 矮人</h1><p>Trigger: <strong><span style="color:rgb(250, 82, 82)">dwarf</span></strong></p><h1 id="heading-816">Gnomes | ノーム | 地精</h1><p>Trigger: <strong><span style="color:rgb(250, 82, 82)">gnome</span></strong></p><h1 id="heading-817">Ogres | オーガ | 巨魔</h1><p>It may be a little overstyled and blurry. You may consider reducing LoRA weight when using ogre trigger.</p><p>Trigger: <strong><span style="color:rgb(250, 82, 82)">ogre</span></strong></p>

Files with SHA256: dd1206c6137e2b5bf689cf93bb102c7e2170c30a581110b55fc376395b28ade4