LoRA model trained on 25 images of anime characters at a black-tie/met gala event.

﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌﹌

For generation, I suggest the following,

Checkpoint: Pony Diffusion V6 XL

Aspect Ratio: Portrait (trained on 1150 × 2052)

Prompt: score_9, score_8_up, score_7_up, source_anime, black-tie event style, detailed background, formal setting, elegant lighting

Negative Prompt: monochrome, 3d, oversaturated

Sampler: DPM++ 2M Karras