Have fun! The entire training dataset is attached to the model:) Recommended Generation Settings<ul><li>Base Models: Photorealistic SDXL checkpoints (RealVisXL V4.0/V5.0, Juggernaut XL, CyberRealistic XL)</li><li>CLIP Skip: <code>1</code> (Crucial for SDXL!)</li><li>CFG Scale: <code>6.0 - 7.0</code></li><li>Aspect Ratio: Strictly vertical formats (<code>768x1024</code> or <code>896x1152</code>, 9:16 ratio) to render the hourglass proportions correctly without stretching.</li></ul>

Files with SHA256: f40d1704a816fc1ad615ce6b422d1332ac315eef3508ee78cd9b0b8f65d69f4d