LoRA is a highly accurate model for illustration that has been trained with AnimaTensor.
Utilizing the learning characteristics of AnimaTensor, it has excellent stability of details, consistency of color, and balance of composition, and it combines a wide range of styles with high resolution.
🌟Features
🖌️ Less line and painting collapse, strong for precise detail depiction
🎞️ Both characters and backgrounds are cohesive, and composition is less likely to break down
🎨 Wide range of styles, flexible use with other LoRA and Checkpoint
⚙️ Stable output even with repeated generation due to AnimaTensor's unique structure
🛠 Recommended environment
✅ Checkpoint: AnimaTensor compatible model (or highly compatible anime CP)
⚖️ LoRA weight: 0.7 to 1.0
🔄 Recommended sampler: Euler. Euler・a
🎛️ Recommended CFG: 4 to 6 (stability emphasized)
💡Application examples
Anime, beautiful girl illustrations, 2.5D-style portraits
Illustrations with backgrounds, cuts with movement and composition
Can be expanded by combining with other LoRA (clothes, expressions, deformations)
AnimaTensor LoRA is capable of producing delicate and high-quality visual expression by taking advantage of the strengths of .
Recommended for creating illustrations that emphasize stability and reproducibility!