This LoRA is trained to replicate the delightful and charming art style of the Japanese artist Mizuki Hitoshi (みずき ひとし).
The style is characterized by its incredibly cute and appealing character designs, vibrant color palettes, soft yet detailed shading, and beautifully rendered expressive eyes. It perfectly captures a "moe" aesthetic that is both sweet and captivating.
Key Features:
Authentic Style: Faithfully reproduces Mizuki Hitoshi's signature look.
Versatile for Characters: Trained primarily on the artist's high-quality illustrations of Touhou Project characters, this model excels at generating fan art of the series. It is also highly adaptable for creating your own Original Characters (OCs) with the same charming aesthetic.
NSFW Potential: Given the artist's background, this model has a good understanding of anatomy and can be guided towards more suggestive or NSFW creations if desired.
How to Use:
Trigger Word: Add mizukihitoshi to your prompt.
Recommended Weight: A weight of 0.7 - 0.9 works best for strong style replication. You can lower the weight to blend the style with other LoRAs or checkpoints.
Helpful Tags: For best results, combine the trigger word with descriptive tags such as anime style, cute, solo, detailed hair, vibrant colors, detailed shading.
Character Generation: To create specific Touhou characters, simply add their name tags (e.g., hakurei reimu, kirisame marisa, remilia scarlet).
Enjoy creating your own adorable characters with this style!