<h2 id="heading-83">Summary</h2><p><span style="color:rgb(236, 236, 236)">This Lora model, named Bailian, is from Wuthering Waves. The SD 1.5 version was trained with a resolution of 514x514. I've also put a lot of focus on training her outfit details to make them as close as possible. Feel free to adjust the settings to match your preferred style. I'll be updating Lora as soon as more images are added to the database</span></p><p></p><h3 id="heading-445">Usage</h3><pre><code>Default Costume Prompt: bare shoulders, black gloves, (dark green hair:1.2), (blue eyes:0.9), black shirt, detached sleeves, earrings, eyeliner, floating hair, gloves, half gloves, jewelry, long bangs, long hair, long sleeves, low-tied long hair, medium breasts, partially fingerless gloves, red eyeliner, shirt, sidelocks, sleeveless shirt, solo, white sleeves</code></pre><p></p><h2 id="heading-84"><span style="color:rgb(250, 82, 82)">Note</span></h2><p><span style="color:rgb(236, 236, 236)">You don't have to use all the additional prompts to generate the costume. Feel free to remove some as needed. I suggest setting the weight between </span><u><span style="color:rgb(236, 236, 236)">0.7</span></u><span style="color:rgb(236, 236, 236)"> and </span><u><span style="color:rgb(236, 236, 236)">1</span></u><span style="color:rgb(236, 236, 236)">, which should work for other costumes as well. If you come across any clothing that can't be removed, lower the weight</span></p>

Bailian | Wuthering Waves

<p>Due to the limited training image database, there are pose issues with Lora, as it was mainly focused on her outfit. This can result in distorted or repeated poses in generated images. For improvement, I suggest using ControlNet and adding more details to Lora for this v.1 version.</p>

Files with SHA256: ff8491e1f32312435b0636b91079581219de3db81f72211af98d3fb8b61ac85d