Trained a lora similar to Illustrious/NoobAI styles. Model itself has a great anime capabilities but mostly it will generate like an anime screencap. This lora help achieving more illustration-like style.

Recommended strength: 0.8-1.3

Trigger word:

elusarca anime style

Recommended prompt prefixes that works well wit Z-Image:

high-quality anime illustration, 2D, clean lineart, cel-shaded, vibrant colors, soft shading

Trained with AI Toolkit by Ostris