Suggested tags for consistent results:Default Outfit - <pre><code>momoi kurba, solo, blonde hair, bob cut, blunt bangs, low-tied sidelocks, red hair bows, pink eyes, pink halo, headphones, black tail, striped tail, white collared shirt, blue necktie, suspenders, oversized white coat, long sleeves, coat draped over shoulders, black pleated skirt, frilled hem, black thighhighs, striped thighhighs, pink boots
</code></pre>Maid Outfit - <pre><code>momoi kurba, solo, blonde hair, bob cut, blunt bangs, sidelocks, twintails, red hair bows, pink eyes, pink halo, headphones, black tail, striped tail, maid, maid headdress, black dress, puffy sleeves, white apron, blue neck ribbon, white pantyhose, mary janes</code></pre>Recommended settings:<ul><li>DPM++ 2M </li><li>Automatic works fine (tested with Karras) </li><li>Sampling Steps: 25-30 </li><li>CFG Scale: 5-6 </li><li>Model Weight: 0.8-1 (I've had good success with weight 0.8) </li><li>Clip Skip: 2 </li></ul>This was the first LoRA that I've trained on multiple outfits. The default outfit and maid are locked in, but I used a total of 10 other outfits for training to help with versatility and body silhouette. As a personal preference, I generally use ADetailer for the faces of any generation, this model slightly blurry eyes without it, so, I would recommend using a face ADetailer.I have had success with multiple Illustrious/XL Checkpoints. However, more stylistic checkpoints (PVC Models for instance) tend to warp and add some details.I would love to see what you create with this LoRA. If you have any feedback, I'd be be interested in hearing it.

Saiba Momoi - [Blue Archive]

First LoRA that I've trained for multiple outfit generation.

Files with SHA256: 845d57e6774cdeb2653d4c3f2f76c406d44c23e16746b659c110fe42f4c193be