This LoRA captures a distinctive high-glam couture style defined by crystal embellishments, liquid shimmer textures, daring cutouts and body-contouring eveningwear silhouettes (inspired by Julien Macdonald). The model learned how to apply sparkling metallic surfaces, deep plunging constructions and provocative glamour styling to a wide variety of garments.It works particularly well for red carpet fashion, high fashion editorial and runway eveningwear, but the style can also transfer surprisingly well to other glamorous or experimental fashion concepts.The training set includes many different types of outfits and couture elements, for example:<ul><li>crystal mesh evening gowns</li><li>shimmering metallic dresses with liquid-like sparkle</li><li>deep plunging necklines and dramatic chest framing</li><li>bold side cutouts and waist-revealing constructions</li><li>body-hugging cocktail dresses and evening gowns</li><li>embellished mini dresses and sculpted midi silhouettes</li><li>sheer illusion elements and sparkling surface textures</li><li>draped glamour dresses with gathered knot details</li><li>provocative red carpet styling</li><li>luxurious runway and editorial fashion looks</li><li></li></ul>The LoRA focuses on materials, surface textures, silhouette language and glamour styling, not on specific identities or faces.Because of this, the style can also transfer well to other fashion domains such as:<ul><li>cocktail dresses</li><li>luxury nightlife looks</li><li>swimwear-inspired glamour concepts</li><li>sensual editorial fashion</li><li>futuristic eveningwear</li><li>experimental cutout fashion</li></ul>This is, just like my other clothing LoRAs, not a "single outfit LoRA". Instead, I trained how fabric works, falls, looks and behaves and how much skin is exposed.Would love if you post your creations here with it.The LoRA was trained on the default flux1-dev model for maximum compatibility.Dataset and training data:<ul><li>94 images</li><li>10 steps per image</li><li>10 epochs</li><li>bucks with 1152px max size, no auto scaling</li><li>cosine scheduler with 0.2 warmup and 0.8 decay</li><li>Captioning was done with OpenAI</li><li>alpha == dim: 32</li></ul>If you use it with other loras, for example a character lora, try to lower the weight - I had good results at around 0.4-0.5.As always: I did not polish any of the images in the showcase, they all came out like this. I did 2 images per prompt and picked the better looking one.<hr />

Files with SHA256: d6839ce4232b7cc61535317dd0accef213f06a461f6cce96e13fa7faa845e69c