<h3 id="heading-2739"><strong><span style="color:#fa5252">Redone</span></strong> my old lora that was pretty bad used different images different training and different base model from the old ones.</h3><h3 id="heading-2740">Was trying out a different way to train lora and decided to take some of my <span style="color:#fab005">bad/failed</span> lora and retry them this version was a result of a few <span style="color:#fab005">attempts</span> with slightly different settings. </h3><h3 id="heading-2741"><span style="color:#12b886">Use at strength 0.5~1.0 tags:</span><span style="color:rgb(18, 184, 134)"> skirt,</span><span style="color:#12b886"> microskirt, miniskirt.</span><br /><em><span style="color:#40c057">use them together or separate but together was the training tokens keep during training.</span></em></h3><h3 id="heading-2742"><span style="color:#e80000">NOT TRAINED ON NAI</span> so it may be ass with certain models.</h3><p></p>

RE-ConceptShortSkirt

<p>normal version no extra settings trained with 53 images </p>

Files with SHA256: 2093893de9af5217f4823260b691d33de2f91469e8089ab39341e13c9ee86240