Trained for 8000 steps using 50 images with 1024*1024 resolution featuring a strong wlop style, with the dataset finely labeled using gemini3.

With 8-step acceleration, it is recommended to set the Lora strength to 1.50, sampler euler, and scheduler simple. The wlop art style is complex and difficult to mimic, so generating multiple images is suggested to achieve the best results.

Compared to the flux version, there is an improvement in single image generation success rate, visual aesthetics, and style similarity. Lora does not require specific prompts, but testing shows that adding keywords like "hand-drawn style," "thick paint," and "brushstrokes" to certain prompts yields better results.