Trigger Word: south-indian-beauty

Ecosystem: Anima (Aesthetic-Focused Diffusion Engine)

Training Engine: AI-Toolkit

Dataset Configuration: 10 Epochs | 1,500 Precision Steps

Recommended Inference Settings

To get the most out of this symmetric Rank 32 Anima LoRA, apply these baseline configurations:

• LoRA Weight: 0.8 - 1.0 (Because of the stable constant learning rate, it holds up beautifully even at a full 1.0 weight).

• Sampling Steps: 25 - 35

• CFG Scale: 3.5 - 5.0

• Sampler: Euler or DPM++ 2M

• Resolution: Best generated at 1024x1024 or native vertical aspect ratios.