A character LoRA of Rias Gremory from High School DxD, trained with a synthetic dataset targeting a anime cel shading aesthetic.
Trigger word: refchar_pilot
Suggested starting strength: 0.6 (used in the saved showcase; adjust to taste).
Training checkpoint: perfectdeliberate_v10.safetensors
For the intended look, try adding: anime style, cel shading, clean line art.
Use the trigger alongside your scene, outfit and composition prompts. Results vary with the checkpoint, prompt and LoRA strength.
Training notes: Experimental release. Trained for 50 steps at 384 x 384, with rank 4. Training checkpoint: perfectdeliberate_v10.safetensors. Character details and style consistency may vary.

