<p><em>(More reliable LoRA model of this character below the description)</em></p><h3 id="heading-27">Version 0.5</h3><ul><li><p>Trained with 20 pictures mainly screenshot from the official game and some official art.</p></li><li><p>I think because lack of background in the dataset made this model generated random background sometimes, so it's recommended to use "<em>(simple background:0.8, plain background:0.8, white background:0.8)" </em>to reduce the random/abstract background.</p></li><li><p>Info about the prompt, model, sampler etc. are in the image above.</p></li></ul><h3 id="heading-28"><s>Version 1.0</s></h3><ul><li><p><s>Soon</s></p></li></ul>

Files with SHA256: 7575d84feffdffd8f4db02fda0e2eccac6ffd097bb298708aff529c242f38e81