<h2 id="lora-based-on-lilac-:-https:x.coma_lil4c">LoRA based on Lilac : <a rel="ugc" href="https://x.com/a_Lil4c">https://x.com/<strong>a_Lil4c</strong></a></h2><h2 id="trained-using-wai-il-v15">Trained using WAI IL V15</h2><h2 id="trigger-words:"><span style="color:rgb(250, 82, 82)">Trigger words:</span></h2><pre><code>lilac stylerx</code></pre><h2 id="recommended-weight:-0.6-0.9"><strong>Recommended weight: 0.6-0.9</strong></h2><p>Note:</p><p>Works well with any body types</p><p>Lower weight will wash away the style a lot, since not much background was trained on this lora and dataset was small, I personally recommend going from at least 0.6 or above as a starting point</p>

[Style | illustrious-XL] Lilac

<p>initial version</p><p>Dataset train was small but high quality, using TRev2 soften most of the issue but do also expect the generated quality to be not as good as my other loras. </p><p>I would treat this as a beta, use with external generator is recommended to give out higher quality results  </p>

Files with SHA256: 256a47db0b303834e1dd9cd4f665eb735390ee54e91f85f7d2cc2be0e00d9411