First version of the LoRA trained on 32 high-quality images of Pinkie Pie from Equestria Girls.

Trained at 1024 resolution, 10 epochs, batch size 5, using Pony Diffusion V6 XL.

This LoRA accurately captures her hair, outfit, expressions, and colors from the series.

Recommended weight: 0.8–1.0 for best results.