<h1 id="iterativel1nes"><span style="color:rgb(76, 110, 245)">IterativeL1nes</span></h1><p>Conversion of one of my personal SDXL mixes into Anima. </p><p>This lora was trained using both tags and natural language prompts. It can respond well to both prompting styles or mixes between them.</p><h3 id="how-to-use"><span style="color:rgb(121, 80, 242)">How to use</span></h3><p>I recommend er_sde + beta/sg_uniform but most commonly used scheduler/sampler combinations with Anima will also work quite well here.</p><p>4-5 CFG recommend / 25 to 30 steps are my general use recommendations.</p><p>The tag @IterL1nes is mainly used for style reinforcement and isn't necessary for the lora to work. All example images are without using @IterL1nes</p>

IterativeLines

<p>2.0 Improved dataset and captioning. This improved the washed out colors present in V1.</p>

Files with SHA256: 5b887acee25d9df53c84e0f6528c5a351bdcbd22e3ecf634ea96bb432790c67f