If you use this LoRA, please consider posting your results in the gallery :( it would help me improve it more easily.

V5

Improved fidelity to the original artwork while significantly enhancing overall image quality. This version features cleaner linework, more consistent facial structure, sharper eyes, richer hair volume/detail, smoother skin rendering, and stronger character identity preservation across generations. Lighting and cel-shading have been refined to keep the classic retro anime feel while delivering a polished modern HD finish. Better pose stability, cleaner anatomy, and more reliable outputs in a wide variety of scenes and outfits. Wide shot scenes have been especially improved, with better long-distance facial detail, stronger character consistency, and cleaner overall scene readability without requiring other LoRAs.

V3

Focused on improving facial fidelity and overall visual consistency.

Delivers more accurate eyes and facial features closer to the original design, with cleaner linework and sharper details.

no additional LoRAs required to achieve strong, high-quality results.

Next improvement goals:

Further development will focus on advanced hair dynamics and interaction

V2

This version is a major improvement over the previous one.
The first model was difficult to use and required very specific prompts to get consistent results.

This new version is much more user-friendly and flexible. It produces cleaner and more stable outputs with simpler prompts, making it easier to generate the Puma Sisters while keeping good accuracy and style consistency.

The model has been optimized for better generalization, improved character recognition, and smoother results overall.

I am still actively working on improving this model, so future updates and refinements are planned.

V1

Not being satisfied with what I could find on the site, I decided to create my own. I'm quite proud of the result